I ,9) is more stable and shows a wider specificity which facilitates the synthesis of varieties of L-amino acids from a-keto acids by the reductive amination reaction.4 ) The nucleotide sequences for the pdh genes from B. sphaericus,9) Thermoactinomyces intermedius, 10) and Rhodococcus Sp.I1) were published. For this paper, we sequenced the pdh gene from B. badius to study the structural relationship from other proteins.The growth condition of B. badius and the cloning of the pdh gene were described previously.2) Plasmids were purified by a plasmid purification kit Qiagen (Qiagen Inc., U.S.A.).The enzyme was purified from B. badius lAM 11059 and Escherichia coli JM109/pBBI 2 ) to homogeneity with a procedure involving ammonium sulfate fractionation, DEAE-Toyopearl, Butyl-Toyopearl, Sephadex G-200, Mono-Q, and Phenyl-Superose HR5/5 column chromatographies. Before the analysis of the NHz-terminal amino acid sequence, the enzyme sample was passed through a TSK Phenyl-5PW column (0.75 x 7.5 em, Tosoh Corp., Japan), and fractionated with a linear gradient of 20-S0% (v/v) acetonitrile containing 0.05% (v/v) trifiuoroacetic acid. The purified PheDH from B. badius (l.Omg (27.SnM» was digested with lysyl-end peptidase (4.2~g (0.14nM), Wako Pure Chemicals, Japan) at 30°C for 16 h. The resulting peptides were separated by HPLC (Waters 600E, U.S.A.), with a reversed-phase column (To soh Corp, Japan, TSK gel ODS80Ts, ¢46 x 150mm). The amino acid sequences of the peptides and that of N-terminus region were identified with an automatic protein sequencer 6625 (MilliGen, U.S.A.).The fragment to be sequenced was subcloned to a plasmid vector pBluescript. A kilo-sequencing kit (Takara Shuzo, Japan) was used to generate shorter clones suitable for sequencing. The sequencing was done by the dideoxy chain termination procedure of Sanger et al., using an automatic DNA sequencer ALF (Pharmacia, Sweden). It was also done with [a-35 S}dCTP as the radioactive label. A Sphl-Clal fragment (1.6 kb) containing the entire pdh gene based on the expression of the enzyme activity, was prepared from the recombinant plasmid pBB 1 2 ) and used for sequence analysis. The nucleotide sequence of the 1.6 kb Sphl-Clal fragment (given in the data base t) showed that there is an ORF encoding the 1140 bp pdh gene ( Fig.) specifying a protein of a calculated molecular weight of 41,350, which is in excellent agreement with the reported value of 41,000 to 42,000 by SDS-PAGE. 2 ) We have purified the enzyme from both of B. badius and the E. coli JMI09 transformant and found that the N-terminal amino acid sequences are identical. The amino acid sequences of some of internal peptides were in good agreement with the predicted amino acid sequence as indicated in Computer alignments by the SWISS-PROT and NBRF-PIR data bases were made by the DNASIS program (Hitachi, Japan). The deduced primary structure of B. badius PheDH is similar to PheDH from B. sphaericus (67.9% identical over 377 a.a.)9) and PheDH from T. intermedius (56.6% over 346 a.a.), 1 0) leucine dehydro...