Several N-alkyl and N,N-dialkylaminomethanesulfonic acids were synthesized (as zwitterions and/or sodium salts) to be tested for utility as biological buffers at lower pH levels than existing Good buffer compounds (aminoalkanesulfonates with a minimum of two carbons between amine and sulfonic acid groups as originally described by Norman Good, and in common use as biological buffers). Our hypothesis was that a shorter carbon chain (one carbon) between the amino and sulfonic acid groups should lower the ammonium ion pK(a) values. The alkylaminomethanesulfonate compounds were synthesized in aqueous solution by reaction of primary or secondary amines with formaldehyde/sodium hydrogensulfite addition compound. The pK(a) values of the ammonium ions of this series of compounds (compared to existing Good buffers) was found to correlate well with the length of the carbon chain between the amino and sulfonate moeties, with a significant decrease in amine basicity in the aminomethanesulfonate compounds (pK(a) decrease of 2 units or more compared to existing Good buffers). An exception was found for the 2-hydroxypiperazine series which shows only a small pK(a) decrease, probably due to the site of protonation in this compound (as confirmed by X-ray crystal structure). X-ray crystallographic structures of two members of the series are reported. Several of these compounds have pK(a) values that would indicate potential utility for buffering at pH levels below the normal physiological range (pK(a) values in the range of 3 to 6 without aqueous solubility problems)- a range that is problematic for currently available Good buffers. Unfortunately, the alkylaminomethanesulfonates were found to degrade (with loss of their buffering ability) at pH levels below the pK(a) value and were unstable at elevated temperature (as when autoclaving)- thus limiting their utility.

Recent preparations of nitrite reductase do not display the heterodimeric quaternary structure obtained previously (total molecular weight 85,000; subunit molecular weights 24,000 and 61,000), but rather yield only the 61,000 molecular weight subunit, even when buffers containing the protease inhibitor phenylmethylsulfonyl fluoride are used. Nevertheless, such preparations retain the high ratio of ferredoxin-linked to methyl viologen-linked enzyme activity which has been previously taken as a characteristic of only the heterodimeric form. These preparations display a siroheme prosthetic group to protein ratio of 1.1. When nitrite reductase samples are frozen during the purification scheme, even though the ferredoxin-linked specific activity does not significantly decrease, enzyme activity-stained native gel electrophoresis of the subsequently purified protein reveals that gels with several bands of activity can be obtained. Further evidence of protein heterogeneity in these preparations comes from N-terminal amino acid analysis which reveals that even nonfrozen preparations contain two major peptides with valine and cysteine as the N-termini. Formation of complexes of purified nitrite reductase with ferredoxin resulted in siroheme difference electronic spectra which resembled those observed previously for monomeric preparations. However, the siroheme midpoint potential of recent preparations of nitrite reductase (-287 mV) is close to that of the heterodimeric preparations. Ultrafiltration studies of crude extracts of the enzyme indicate that, at least at certain stages of the preparation, higher molecular weight forms of the enzyme may exist. We conclude that the 24,000 molecular weight polypeptide is a contaminant and that the heterodimeric quaternary structure model for spinach nitrite reductase is incorrect. Furthermore, the monomeric preparations we do obtain display both significant protein heterogeneity and facile loss of siroheme upon gel filtration.

We used site-directed mutagenesis to introduce both a NdeI restriction endonuclease site and an initiator codon at the junction of the leader and structural gene sequences of the metallo-beta-lactamase of Bacillus cereus 5/B/6. This construct allowed us to clone just the beta-lactamase structural gene sequence into an Escherichia coli expression vector. E. coli cells were transformed with the recombinant plasmid, the B. cereus beta-lactamase was expressed, and these E. coli cells were disrupted by sonic oscillation. When the resultant suspensions were clarified by ultracentrifugation, the B. cereus beta-lactamase represented 15% of the total protein in the supernatant. Subsequent gel filtration and ion-exchange chromatography allowed the first reported purification to homogeneity of the B. cereus beta-lactamase from E. coli with an 87% recovery and an overall yield of 17 mg of enzyme per liter of cell culture. The electrophoretic mobilities of the enzyme expressed in and purified from E. coli and the enzyme purified directly from B. cereus were identical in both native and sodium dodecyl sulfate gel electrophoreses. As with the B. cereus enzyme, Km and Vmax (using cephalosporin C as substrate) for the enzyme purified from E. coli were 0.39 mM and 1333 units/mg protein, respectively. Likewise, the Co(II)-reconstituted enzyme purified from E. coli, which retained 29% of the activity of the Zn(II) enzyme, had electronic absorption spectra with maxima at 347, 551, 617, and 646 nm with extinction coefficients of 900, 250, 173, and 150 M-1 cm-1, respectively.