Specialized transduction of D-serine deaminase genes: formation of lysogens that yield high lambda-d dsd/lambda ratios and formation of a dimeric lambda-d dsd

In vivo D-serine deaminase transcription start sites in wild-type Escherichia coli and in dsdA promoter mutants

The D-serine deaminase structural (dsdA) and regulatory (dsdC) genes are transcribed with opposite polarity from an intergenic region comprising more than 600 base pairs. The order of genes in the dsd region is supN-dsdA-dsdC-aroC---his. The DNA sequence of the intergenic region has been slightly revised from a previously published version (E. McFall and L. Runkel, J. Bacteriol. 154:1508-1512, 1983). The dsdA gene is preceded by a long open reading frame. The dsdA in vivo transcription start sites for the wild type (base pair +1) and for three phenotypically distinct promoter constitutive mutants were determined by the S1 nuclease method. They are identical and are located about 81 base pairs upstream of the translation start site. D-Serine deaminase regulation is normal in rho mutants. Possible mechanisms for dsdA activation are discussed.

Cloning and expression of the metE gene in Escherichia coli

A lambda-transducing phage was isolated that contains the metE gene. This gene codes for N5-methyl-H4-folate:homocysteine methyltransferase (EC 2.1.1.14), an enzyme that catalyzes the terminal reaction in methionine biosynthesis. A 9.1-kb EcoR1 fragment of this phage, containing the metE gene, was then cloned into pBR325. This plasmid, pJ19, was used to transform Escherichia coli strain 2276, a metE mutant, and restore the MetE+ phenotype. Although the transformed cells produced large amounts of the metE protein in vivo, in vitro studies using pJ19 as template showed low synthesis of the metE protein.

Purification and characterization of D-serine deaminase activator protein

We purified the dsdC gene product, the specific activator of dsdA (D-serine deaminase) gene expression, to about 25% homogeneity from a strain in which its expression was amplified 100-fold. The purification involved, successively: DNase and high-salt treatment of cell extracts, DNA-cellulose chromatography, and Dyematrex (Amicon Corp.) column chromatography. We identified the protein as a discrete spot on two-dimensional O'Farrell gels after the DNA-cellulose step and quantitated it by densitometry. The active form was found to be a dimer. We estimated that there were eight activator dimers per wild-type cell. The activator is a slightly basic protein, with an experimental Km for its ligand D-serine of about 7 X 10(-6)M. The low concentration of the activator in wild-type cells and its autorepression may explain the previously observed partial dominance of dsdC+ in dsdCc/dsdC+ merodiploids.

DNA sequences of the D-serine deaminase control region and N-terminal portion of the structural gene

We determined the DNA sequence of the D-serine deaminase promoter region and of the N-terminal region of the structural gene. There are possibilities in the promoter for secondary structure and for initiation recognition sequences, and there is an open reading frame. The N-terminal sequence for the structural gene confirms that part of the amino acid sequence previously determined by E. Schlitz and W. Schmitt (FEBS Lett. 134:57-62, 1981), including the active site of the enzyme, and spans the two regions unresolved by their work.

Identification and control of synthesis of the dsdC activator protein

Operon fusions between the D-serine deaminase regulatory and structural genes and lacZ were constructed and used to examine the control of expression of the positive regulatory gene, dsdC. Merodiploid strains containing both dsdCp::Mu d (lac Apr) and dsdC+A+ produced only one-fourth as much beta-galactosidase as did the haploid dsdCp::Mu d (lac Apr) strains, indicating that the dsdC+ product repressed its own synthesis. The repression was reversed by D-serine. dsdC expression was not depressed in a cya background. The basal level of D-serine deaminase was the same in wild-type and dsdCp fusion strains. The dsdC gene product was identified in maxicell strains harboring dsd plasmids as a 34,000-dalton protein. dsdC gene transcription proceeded clockwise; thus, its promoter is adjacent to that of dsdA.

Physical mapping of the Escherichia coli D-serine deaminase region: contiguity of the dsd structural and regulatory genes

The genes dsdA, dsdO, and dsdC have been located on a 3.0-kilobase pair (kb) fragment of the Escherichia coli chromosome by a combination of techniques. The loci were first cloned onto lambda and various plasmid vectors. dsd hybrid plasmids were then digested with restriction enzymes, and the fragments were recloned to test for the presence of dsdC or dsdA. In one case, a 4.2-kb restriction fragment containing the dsdA operon was used to form a heteroduplex with a well-defined lambda dsd deoxyribonucleic acid. The results show that dsdA, dsdO, and at least 0.6 kb of dsdC are present on this piece of deoxyribonucleic acid. On the basis of the mapping analysis and the molecular weight of D-serine deaminase, 1.9 kb of the 4.2-kb fragment is accounted for by the three dsd loci. We conclude that dsdO and dsdC are contiguous. A detailed dsd restriction map is presented.

Role of the dsdC activator in regulation of D-serine deaminase synthesis

The activator of the D-serine deaminase operon, the product of the dsdC gene, has been partially purified. It is reasonably stable to routine purification procedures in the presence of its ligand D-serine, but not in its absence. It loses activity upon dialysis in amino acid-free buffer, but activity is completely restored upon readdition of D-serine. It apparently functions purely as an activator, no repressor function could be demonstrated at suboptimal D-serine concentration. It is a transcriptional control element. The time required for in vitro transcription of D-serine deaminase mRNA, nearly 4 min, is similar to that for beta-galactosidase. Since the beta-galactosidase monomer is a much protein, this is surprisingly long.

Role of small molecules in regulation of D-serine deaminase synthesis

Cyclic AMP is required for optimal synthesis of D-serine deaminase synthesis from dsdO+ templates and for optimal hyperinducible synthesis from low constitutive dsdO templates both in vitro and in vivo. Neither D-serine, cyclic AMP, nor dsdC activator has an effect on expression of a high constitutive dsdO template. The synthesis of the dsdC activator itself in vitro is independent of cyclic AMP. Guanosine tetraphosphate does not have a significant effect on in vitro D-serine deaminase synthesis from dsdO+ or dsdO templates. A previously described class of dsdO mutants showing partial catabolite sensitivity of constitutive D-serine deaminase synthesis proved to be low dsdO types. They all contain a low constitutive dsdC mutation; the two effects are additive with regard to level of constitutivity, but only that portion of synthesis attributable to the dsdC mutation is cyclic AMP dependent.

Positive control of D-serine deaminase synthesis in vitro

Efficient constitutive synthesis of D-serine deaminase [D-serine hydro-lyase (deaminating); EC 4.2.1.14] is obtained in vitro by using a slightly modified Zubay system programmed with dsdO6 dsdA+DNA. Synthesis from a dsdO+ dsdA+ template requires active dsdC gene product and 3':5'-cyclic AMP. In vitro synthesis of dsdC product is obtained with a dsdC+ dsdO+ dsdA+ or a dsdCc dsdO+ dsdA+ template; this synthesis is thermosensitive and can be uncoupled from D-serine deaminase synthesis by temperature shift.

Escape synthesis of D-serine deaminase in lambda dsdC integration lysogens

Heat shock of lysogens that contain lambda thermosensitive repressor mutants integrated into the dsdC gene results in escape synthesis of d-serine deaminase at a high differential rate.