By site-directed mutagenesis, TEM-1 beta-lactamase was altered to contain single amino acid changes of E104K, R164S, and E240K, in addition to double changes of E104K/R164S or R164S/E240K and the triple change of E104K/R164S/E240K. Hydrolysis rates for cephaloridine and benzylpenicillin were lowered at least 1 order of magnitude for all enzymes containing R164S substitutions. All mutant enzymes exhibited increased kcat values for beta-lactam antibiotics containing an aminothiazole oxime side chain. Hydrolysis of ceftazidime was most affected, with kcat values increased 3-4 orders of magnitude in all enzymes with the substituted R164S moiety. Km values decreased for all substrates except ceftazidime in the enzymes with multiple mutations. Aztreonam was most affected, with Km values lowered 23-56-fold in the enzymes bearing multiple mutations. When the crystal structures of aztreonam and related monobactams were studied and projected into an active-site model of the PC1 beta-lactamase, it became apparent that the two lysine residues might serve equivalent roles by interacting with the carboxylate of the aminothiazole oxime side chain. Hydrogen-bonding interactions involving the oxime and N7 of the lysine, particularly Lys-104, may also be important in some antibiotics. Ser-164 apparently serves an indirect role, since it is somewhat distant from the active-site cleft.

Growth of a hybridoma culture, along with production of monoclonal antibody, was demonstrated over extended periods in polysulfone hollow fiber membrane modules. The molecular weight cutoffs of the membranes were 70,000, 50,000, and 100,000 daltons. The hybridoma cell line, designated 65/26, produced IgG (2b/kappa) directed at mouse thymus cell surface antigen, TL.1. Cell growth occurred in the shell space of the reactor, using supplemented RPMI 1640 (20% fetal bovine serum) supplied from a separate reservoir vessel through the hollow fiber lumen. The reservoir contained 125 mL media, which was changed every 4 days. Concentrations of immunoglobulin were determined by an enzyme immunoassay (using protein A and alkaline phosphatase-labeled antibody conjugate). For the 10K, 50K, and 100K hollow fiber membrane modules, the maximum IgG concentrations detected in the 2.5-mL shell space were 47.5-80, 510, and 740 mug/mL, respectively. In the 125-mL reservoir for the 100K hollow fiber membrane module, the IgG concentration was measured at 260 mug/mL These values compare with an IgG concentration of 1 mug/mL when grown in a standard tissue culture flask and 3.2-7.6 mug/mL when grown in 100 ml media in a spinner flask. In addition, 10K and 50K hollow fiber membrane modules were run in a mode that decreased the fetal bovine serum supplement with time. Differences between these systems suggest that it is possible to obtain high IgG accumulation rates, both during and after the exponential growth phase of the hybridoma population.