CLONING, SEQUENCING, AND MOLECULAR ANALYSIS OF THE ACETOACETATE DECARBOXYLASE GENE REGION FROM CLOSTRIDIUM-ACETOBUTYLICUM

Acetoacetate decarboxylase (ADC) (EC 4.1.1.4) of Clostridium acetobutylicum DSM 792 was purified to homogeneity, and its first 25 N-terminal amino acids were determined. Oligonucleotide probes deduced from this sequence were used to detect positive clones in partial gene banks derived from Sau3A and HaeIII digests with following ligation into the vector pUC9. In Escherichia coli, the 2.1-kbp HaeIII clones expressed high levels of ADC activity. The expression was independent of the orientation of the insert with respect to the lac promoter of the vector and also of the addition of isopropyl-β-D-thiogalactopyranoside, thus indicating that sequences located on the clostridial DNA controlled transcription and translation. From the E. coli clone with the recombinant plasmid pUG93 containing the 2.1-kbp HaeIII fragment, the ADC protein was purified and compared with the native enzyme. Both were indistinguishable with respect to the molecular mass of subunits and native protein as well as to activity stain. The 2.9-kbp Sau3A fragment could be shown to contain the amino terminus of the acetoacetate decarboxylase (adc) gene but did not express enzyme activity. It partially overlapped with the HaeIII fragment, spanning together 4,053 bp of the clostridial genome that were completely sequenced. Four open reading frames (ORFs) could be detected, one of which was unambiguously assigned to the acetoacetate decarboxylase (adc) gene. Amino acid sequences of the N terminus and the catalytic center as deduced from the nucleotide sequence were identical to sequences obtained from direct analysis of the protein. Typical procaryotic transcriptional and translational start and stop signals could be found in the DNA sequence. Together with these regulatory sequences, the adc gene formed a single operon. The carboxyl terminus of the enzyme proved to be rather hydrophobic. In vitro transcription-translation assays resulted in formation of ADC and ORF3 gene product; the other two ORFs were not expressed. Whereas no homology of the adc gene and ORF2 could be detected with sequences available in the EMBL or GenBank data bases, the obviously truncated ORF1 showed significant similarity to α-amylase of Bacillus subtilis. The restriction pattern and N-terminal amino acid sequence (as deduced from the nucleotide sequence) of ORF3 proved to be identical to those of the large subunit of acetoacetyl coenzyme A:acetate/butyrate:coenzyme A transferase.