The emerging periplasm-localized subclass of AroQ chorismate mutases, exemplified by those from Salmonella typhimurium and Pseudomonas aeruginosa.

BACKGROUND: Chorismate mutases of the AroQ homology class are widespread in the Bacteria and the Archaea. Many of these exist as domains that are fused with other aromatic-pathway catalytic domains. Among the monofunctional AroQ proteins, that from Erwinia herbicola was previously shown to have a cleavable signal peptide and located in the periplasmic compartment. Whether or not this might be unique to E. herbicola was unknown. RESULTS: The gene coding for the AroQ protein was cloned from Salmonella typhimurium, and the AroQ protein purified from both S. typhimurium and Pseudomonas aeruginosa was shown to have a periplasmic location. The periplasmic chorismate mutases (denoted *AroQ) are shown to be a distinct subclass of AroQ, being about twice the size of cytoplasmic AroQ proteins. The increased size is due to a carboxy-terminal extension of unknown function. In addition, a so-far novel aromatic aminotransferase was shown to be present in the periplasm of P. aeruginosa. CONCLUSIONS: Our analysis has detected a number of additional *aroQ genes. The joint presence of *AroQ, cyclohexadienyl dehydratase and aromatic aminotransferase in the periplasmic compartment of P. aeruginosa comprises a complete chorismate-to-phenylalanine pathway and accounts for the "hidden overflow pathway" to phenylalanine described previously.

Expression and characterization of glycosylated and catalytically active recombinant human alpha-galactosidase A produced in Pichia pastoris.

Fabry disease is an X-linked inborn error of glycolipid metabolism caused by deficiency of the lysosomal enzyme alpha-galactosidase A. This enzyme is responsible for the hydrolysis of terminal alpha-galactoside linkages in various glycolipids. An improved method of production of recombinant alpha-galactosidase A for use in humans is needed in order to develop new approaches for enzyme therapy. Human alpha-galactosidase A for use in enzyme therapy has previously been obtained from human sources and from recombinant clones derived from human cells, CHO cells, and insect cells. In this report we describe the construction of clones of the methylotrophic yeast Pichia pastoris that produce recombinant human alpha-galactosidase A. Recombinant human alpha-galactosidase A is secreted by these Pichia clones and the level of production is more than 30-fold greater than that of previously used methods. Production was optimized using variations in temperature, pH, cDNA copy number, and other variables using shake flasks and a bioreactor. Expression of the human enzyme increased with increasing cDNA copy number at 25 degrees C, but not at the standard growth temperature of 30 degrees C. The recombinant alpha-galactosidase A was purified to homogeneity using ion exchange (POROS 20 CM, POROS 20 HQ) and hydrophobic (Toso-ether, Toso-butyl) chromatography with a BioCAD HPLC Workstation. Purified recombinant alpha-galactosidase A was taken up by fibroblasts derived from Fabry disease patients and normal enzyme levels could be restored under these conditions. Analysis of the carbohydrate present on the recombinant enzyme indicated the predominant presence of N-linked high-mannose structures rather than complex carbohydrates.

Purification and characterization of human alpha-galactosidase A expressed in insect cells using a baculovirus vector.

Fabry disease is an X-linked inborn error of glycolipid metabolism caused by deficiency of the lysosomal enzyme alpha-galactosidase A. The enzyme is responsible for the hydrolysis of terminal alpha-galactoside linkages in various glycolipids. To perform more extensive biochemical characterization and to develop new approaches for enzyme therapy, a method of producing and purifying recombinant alpha-galactosidase A suitable for scale-up manufacture for use in humans is needed. Previously, a catalytically active recombinant human alpha-galactosidase A was expressed using a baculovirus vector and purified using conventional chromatography. However, the level of expression was too low to permit economical production and the chromatographic techniques used for enzyme purification were not suitable for enzyme to be used in humans. Therefore, the cDNA of the enzyme was cloned to an improved baculovirus vector and the enzyme was expressed in a 15-liter bioreactor using optimized growth conditions. Infection of insect cells by the baculovirus resulted in a significant fivefold increase in the level of secreted recombinant alpha-galactosidase A activity that is compatible with economic manufacturing. The recombinant alpha-galactosidase A was purified to homogeneity using ion exchange (Poros 20-CM, Poros 20-HQ) and hydrophobic chromatography (Toso-ether, Toso-butyl) using the BioCAD HPLC workstation. These chromatographic steps are readily scalable to larger volumes and are appropriate for the purification of the recombinant human alpha-galactosidase A to be used in clinical trials of enzyme replacement therapy for Fabry disease patients.

Urease from a potentially pathogenic coccoid isolate: purification, characterization, and comparison to other microbial ureases.

Strain SL100 is a gram-positive coccoid isolate prototype with an adhesin specific for gastric mucin and is representative of potentially pathogenic organisms obtained at biopsy from patients with gastric disorders. The urease of this isolate constitutes a significant fraction of the total cell protein, and the outcome of the purification strategy described herein suggests that it is associated with a cell wall fraction. The urease was purified 138-fold to apparent homogeneity, as indicated by gel electrophoresis, to a specific activity of 1,120 U/mg. The urease was unstable during purification in the absence of nickel, which is present in a metallocenter in other microbial ureases. When nickel sulfate was present during growth (5 microM) and in buffers during sonication and purification (100 microM), the urease was completely stable at room temperature during the purification procedure. The native urease was approximately 260 kDa and was composed of three subunits of 65 kDa and three subunits of 21 kDa. The purified urease was relatively stable in acid and retained most of its activity after incubation for 30 min at pH 1.3. The K(m)s for urease measured from whole cells and for the purified enzyme were 0.56 and 1.7 mM, respectively, indicating that some cell wall component(s) affects the affinity of the enzyme for urea. The V(max)s for urea hydrolysis measured from whole cells and for the purified enzyme were 8.1 and 1,120 mol/min/mg of protein, respectively. The kinetic parameters, relative abundance, and subunit composition are more similar to those of the ureases of Helicobacter than to those of the ureases of other microbial species. These similarities are consistent with an adaptation of this organism to colonization of the stomach and indicate that the urease may be a virulence factor during colonization.

Inhibition of recombinant Pneumocystis carinii dihydropteroate synthetase by sulfa drugs.

Forty-four sulfa drugs were screened against crude preparations of recombinant Pneumocystis carinii dihydropteroate synthetase. The apparent Michaelis-Menten constants (Km) for p-aminobenzoic acid and 7,8-dihydro-6-hydroxymethylpterin pyrophosphate were 0.34 +/- 0.02 and 2.50 +/- 0.71 microM, respectively. Several sulfa drugs, including sulfathiazole, sulfachlorpyridazine, sulfamethoxypyridazine, and sulfathiourea, inhibited dihydropteroate synthetase approximately as well as sulfamethoxazole, as determined by the concentrations which cause 50% inhibition and/or by Ki. For all sulfones and sulfonamides tested, unsubstituted p-amino groups were necessary for activity, and sulfonamides containing an N1-heterocyclic substituent were found to be the most effective inhibitors. Folate biosynthesis in isolated intact P. carinii was approximately equally sensitive to inhibition by sulfamethoxazole, sulfachlorpyridazine, sulfamethoxypyridazine, sulfisoxazole, and sulfathiazole. Two of these drugs, sulfamethoxypyridazine and sulfisoxazole, are known to be less toxic than sulfamethoxazole and should be further evaluated for the treatment of P. carinii pneumonia.

Characterization of glycosylated and catalytically active recombinant human alpha-galactosidase A using a baculovirus vector.

Fabry disease is an X-linked inborn error of glycolipid metabolism caused by a deficiency of the lysosomal enzyme alpha-galactosidase A (GalA; EC 3.2.1.22). In order to obtain large quantities of this human enzyme for physical characterization and for the development of new approaches for enzyme therapy, we constructed derivatives of the Autographa californica nuclear polyhedrosis virus that produce the human enzyme. The recombinant GalA (re-GalA) is produced at high levels, and is active with both the artificial substrate, 4-methylumbelliferyl-alpha-D-galactopyranoside, and the natural in vivo substrate, trihexosylceramide. The purified re-GalA is glycosylated and is taken up by normal and Fabry fibroblasts in cell culture. Mass spectral analysis of total monosaccharides released by hydrazinolysis indicates that it contains fucose, galactose, mannose and N-acetylglucosamine. Amino-acid sequence analysis of six proteolytic peptides corresponded to sequences predicted by the cDNA. The molecular masses of the purified enzyme, estimated by electrospray mass spectroscopy and laser desorption time-of-flight analysis are 46.85 and 46.62 kDa, respectively, approx. 10% greater than the polypeptide portion predicted by the cDNA. The recombinant enzyme retains significant catalytic activity after modification with poly(ethylene glycol), a treatment which decreases the immunogenicity and increases the circulation life of many proteins used therapeutically.

Multiple transcripts encoded by the ilvGMEDA gene cluster of Escherichia coli K-12.

We report here that, using Northern (RNA) blots, we identified two relatively stable transcripts of 4.6 and 1.1 kb that correspond to the products of the ilvEDA and ilvE genes and two relatively unstable transcripts of 6.7 and 3.6 kb that correspond to the products of the ilvGMEDA and ilvDA genes. The transcripts were identified by the use of eight probes derived from segments of the ilvGMEDA cluster. In addition, we used two strains with deletions of ilvG or ilvDA and observed the expected decrease in transcript size in Northern blots. Primer extension with reverse transcriptase generated a 169-nucleotide product corresponding to a 5' end within the ilvED intercistronic region, 37 nucleotides from the AUG codon of the ilvD gene. This primer extension product presumably indicates the 5' end of the ilvDA transcript that we detected in Northern blots. The stability of the transcripts was monitored, and RNase E was found to play a major role in ilv transcript degradation. Transcript levels varied in response to growth in the presence of the end product amino acids and in response to the presence of the polar frameshift site in ilvG. Although there have been speculations about the identities and numbers of transcripts derived from the ilvGMEDA cluster on the basis of the identification of some of the sites of transcription initiation and termination, this is the first report of the use of Northern blots to determine the actual sizes and distribution of mRNAs present in vivo.

Sequence and transcriptional activity of the Escherichia coli K-12 chromosome region between rrnC and ilvGMEDA.

We previously identified a protein related to the expression of the ilvGMEDA cluster of Escherichia coli K-12. It was observed that this ilv-related protein was produced at higher levels in UV irradiated cells infected with lambda dilvGMEDA phage with specific ilvG mutations (ValR), compared to phage carrying the wild-type(ValS) ilvG allele. The gene encoding this protein was further localized to a region between rrnC and ilvGMEDA by analyzing restriction fragment subsets in maxicells. We have now determined the nucleotide (nt) sequence of the 3.5-kb segment between rrnC and ilvGMEDA, and two open reading frames (ORFs) are present in the region expected to contain the ilv-related gene. These ORFs predicts Mrs of 18,751 (ORFI) and 20,085 (ORFII) Da, and both ORFs have a strong probability to encode proteins based on codon frequency analysis. Maxicell analysis revealed that a 1319-bp HindIII-SmaI fragment containing ORFI encodes the ilv-related peptide. We deleted a ClaI fragment that removed a portion of ORFI encoding the C-terminal region of the peptide, and maxicell analysis revealed a decrease in the size of the protein produced in accord with the prediction. RNA slot blots and Northern blots were used to characterize transcripts encoding ORFI. A transcript initiated 112 nt from the ilvGp2 promoter, but proceeding in the opposite direction, may encode the ORFI peptide.

The complete nucleotide sequence of the ilvGMEDA cluster of Escherichia coli K-12.

The ilvGMEDA gene cluster of Escherichia coli K-12 has been the focus of intensive genetic and biochemical analysis for the past 30 years. Genetic regulation of the ilvGMEDA cluster involves attenuation, internal promoters, internal Rho-dependent termination sites, a site of polarity in the ilvG pseudogene of the wild-type organism, and autoregulation by the ilvA gene product, the biosynthetic L-threonine deaminase. We have now completed the nucleotide sequence of the 6600-bp cluster and have analyzed it, along with the ilvYC, ilvBN, and ilvIH genes, for codon frequencies and possible evolutionary relationships. The isoleucine content of each of the gene products of the ilvGMEDA cluster is quite similar (less than a two-fold variation), thus excluding one possible interpretation of the isoleucine-specific downstream amplification phenomenon. There is no evidence for retrograde evolution in the cluster since no significant homologies are detectable among genes that catalyze sequential reactions of the pathway. A highly significant homology does exist, however, between the threonine deaminases of yeast mitochondria and E. coli. The sequence at the boundary of the ilvA and ilvD genes is TAATAATG, so that the second TAA stop codon of ilvD overlaps the ATG initiation codon of ilvA.

Expression of the human alpha-galactosidase A in Escherichia coli K-12.

We used the prokaryotic expression vector, ptrpL1, for the expression in Escherichia coli K-12 of a cDNA clone specific for the human lysosomal hydrolase, alpha-galactosidase A. The 5' terminus of the cDNA clone was engineered so that an ATG codon precedes the first codon of the mature form of the enzyme. A clone with elevated expression of this human enzyme was constructed by increasing the distance between the Shine-Dalgarno site and the ATG start codon from 6 to 8 bp. Clones with alpha-galactosidase A specific cDNA encoding the proenzyme produce a protein of 45 kDa, the size expected for the intact proenzyme. The 45-kDa protein is specifically precipitated by antibody to alpha-galactosidase A, and its expression is repressed by tryptophan and induced by 3-beta-indoleacrylic acid as expected for this expression vector. The human enzyme is produced in E. coli in a catalytically active form at levels sufficient to support the growth of cells using alpha-galactosides as sole sources of carbon and energy. In addition, bacterial colonies that produce the human enzyme turn blue in the presence of 5-bromo-4-chloro-3-indolyl-alpha-D-galactopyranoside.

A genomic clone containing the promoter for the gene encoding the human lysosomal enzyme, alpha-galactosidase A.

We have isolated and characterized a human genomic clone for a lysosomal enzyme gene. The start point of transcription was identified using primer extension of poly(A)+ mRNA. This genomic clone is specific for human alpha-galactosidase A, and it includes sequences for the promoter, complete signal peptide, first exon, and part of the first intron. Direct and inverted repeat elements of 10, 11, 16, 19, and 22 nucleotides (nt) flank the promoter site. A (GA)n repeat element of approx. 60 nt with strong homology to similar elements identified in several species is located upstream from the promoter. A GGGCGG site specific for DNA-binding protein Sp1 is located near a CAAT box, and the CCGCCC inverted repeat of the Sp1 binding sequence is located by the TATA box. The sequence immediately flanking the ATG start codon of the human alpha-galactosidase A is highly homologous to sequences flanking the ATG start codons of the other human lysosomal hydrolases for which sequence information is available (beta-glucocerebrosidase, cathepsin B, cathepsin D, and beta-hexosaminidase alpha chain), but not for any of the other 133 human signal peptides examined. Our analysis also reveals that conversion of the propeptide to the mature enzyme involves cleavage of a C-terminal rather than an N-terminal fragment. This information about the normal alpha-galactosidase A gene will be useful for comparison to data obtained from patients with Fabry disease, who are characterized by a deficiency of this enzyme. This is the first genomic clone described to date for any lysosomal enzyme, and it establishes a reference for future analyses of the molecular events that mediate the expression of this important class of enzymes.

Human alpha-galactosidase A: nucleotide sequence of a cDNA clone encoding the mature enzyme.

The complete nucleotide sequence has been determined for a lambda gt11 cDNA clone (lambda AG18) containing the full-length coding region for the mature lysosomal form of human alpha-galactosidase A (alpha-Gal A; EC 3.2.1.22). The lambda AG18 insert contained a 1226-base-pair sequence with an open reading frame encoding 398 amino acids of the mature polypeptide (predicted Mr = 45,356) and the last 5 amino acids of the propeptide sequence. The poly(A) signals AATACA and ATTAAA occurred 28 and 11 nucleotides prior to the TAA stop codon, respectively. There was no 3' untranslated region as the poly(A) sequence immediately followed the TAA termination codon; a second independently cloned cDNA confirmed this finding. The predicted amino acid sequence was colinear with 86 nonoverlapping residues (22% of the mature subunit) determined by microsequencing amino-terminal, tryptic, and cyanogen bromide peptides of the purified mature enzyme. Four potential N-glycosylation sites were identified, all of which occurred at predicted beta turns in hydrophilic regions of secondary structure. RNA transfer hybridization analysis of HeLa poly(A)+ RNA demonstrated a single 1.45-kilobase band whose signal was decreased by prior immunoabsorption of polysomes with monospecific alpha-Gal A antibodies. Searches of nucleic acid and protein data bases did not reveal significant homology even with the limited sequences available for mammalian lysosomal enzymes.

Isolation of a cDNA clone for the human lysosomal proteinase cathepsin B.

The cysteine proteinase cathepsin B is one member of the lysosomal acid hydrolases. Based on the peptide sequence of rat liver cathepsin B, an oligonucleotide mixture containing 128 different 17-mers was synthesized and used as a probe to screen adult and fetal human liver cDNA libraries. A recombinant clone with a 1540-nucleotide insert was identified from the fetal library, and DNA sequence analysis confirmed that this clone encodes human cathepsin B. The clone, designated pCB-1, has sequences for 81% of the coding region (for amino acid residues 50-252) together with approximately equal to 880 nucleotides of the 3' untranslated region of the mRNA. The DNA sequence also shows that the predicted carboxyl terminus of the coding sequence is longer than the mature protein by 6 amino acid residues. Southern blot analysis of restriction enzyme digests of human placental DNA revealed a simple pattern of hybridizing fragments using the cathepsin B coding sequence as probe. The result suggests that there is a single copy of cathepsin B gene per haploid genome.

Fabry disease: isolation of a cDNA clone encoding human alpha-galactosidase A.

Fabry disease is an X-linked inborn error of metabolism resulting from the deficient activity of the lysosomal hydrolase, alpha-galactosidase A (alpha-Gal A; alpha-D-galactoside galactohydrolase, EC 3.2.1.22). To investigate the structure, organization, and expression of alpha-Gal A, as well as the nature of mutations in Fabry disease, a clone encoding human alpha-Gal A was isolated from a lambda gt11 human liver cDNA expression library. To facilitate screening, an improved affinity purification procedure was used to obtain sufficient homogeneous enzyme for production of monospecific antibodies and for amino-terminal and peptide microsequencing. On the basis of an amino-terminal sequence of 24 residues, two sets of oligonucleotide mixtures were synthesized corresponding to adjacent, but not overlapping, amino acid sequences. In addition, an oligonucleotide mixture was synthesized based on a sequence derived from an alpha-Gal A internal tryptic peptide isolated by reversed-phase HPLC. Four positive clones were initially identified by antibody screening of 1.4 X 10(7) plaques. Of these, only one clone (designated lambda AG18) demonstrated both antibody binding specificity by competition studies using homogeneous enzyme and specific hybridization to synthetic oligonucleotide mixtures corresponding to amino-terminal and internal amino acid sequences. Nucleotide sequencing of the 5' end of the 1250-base-pair EcoRI insert of clone lambda AG18 revealed an exact correspondence between the predicted and known amino-terminal amino acid sequence. The insert of clone lambda AG18 appears to contain the full-length coding region of the processed, enzymatically active alpha-Gal A, as well as sequences coding for five amino acids of the amino-terminal propeptide, which is posttranslationally cleaved during enzyme maturation.

Assessment of plasmid profile, exoenzyme activity, and virulence in recent human isolates of Yersinia enterocolitica.

We examined a group of 23 recent clinical isolates of Yersinia enterocolitica recovered from symptomatic patients residing in the New York, N.Y. area. These isolates were tested for the presence of plasmids, exoenzyme activity, mouse lethality, and phenotypic properties postulated to correlate with virulence. Of the 23 isolates, 17 harbored a 60- to 65-kilobase (kb) plasmid. Six isolates were lethal for white mice, showed the phenotypic markers of autoagglutination and calcium dependence for growth at 37 degrees C, and contained a 60- to 65-kb plasmid. Restriction endonuclease analysis with several different enzymes revealed the presence of three distinct plasmid profiles in these isolates. Isolates with a single plasmid of 60 to 70 kb, typical for this species, were detected, but these were of three distinct types as judged from restriction enzyme digestion. One strain was unusual among clinical isolates of Y. enterocolitica in that it contained at least four distinct plasmids. In addition, this nontypable strain showed exoenzymatic activity similar to that of serogroup O8 isolates, was not lethal to mice, and did not require calcium for growth at 37 degrees C.

Internal promoter in the ilvGEDA transcription unit of Escherichia coli K-12.

Segments of the ilvGEDA transcription unit have been cloned into the promoter tester plasmid pMC81. This vector contains cloning sites situated upstream of the lacZ gene coding for beta-galactosidase. Using this method we have quantitatively evaluated in vivo (i) the activity of previously described promoter, pG, preceding ilvG; (ii) the relative activity of pE promoter, previously postulated to be located between ilvG and ilvE; and (iii) the effect of the frameshift site present in the wild-type ilvG gene by comparison with mutant derivatives lacking this frameshift site. Isogenic derivatives of strain MC1000 were constructed by transduction with phage P1 grown on rho-120, delta(ilvGEDA), delta(ilvED), and ilvA538 hosts. The potential effects of these alleles that were previously postulated to affect ilvGEDA expression were assessed in vivo by monitoring beta-galactosidase production directed by ilv DNA fragments. Cloned ilv segments were also tested for activity in vitro with a DNA-directed coupled transcription and translation system. The production in vitro of ilv-directed ilv gene expression and beta-galactosidase expression with ara-ilv-lac fusions paralleled the in vivo activity.

Location of the rho gene and characterization of lambda ilv-gal derivatives of lambda ilv-rho bacteriophage.

The location of the rho gene and its position relative to the ilv genes of Escherichia coli K-12 was analyzed using genetic criteria, restriction enzyme cleavage, and maxicell analysis. Plasmids were constructed with deletions of the rho gene introduced in vitro, and lambda ilv-gal derivatives of lambda ilv-rho bacteriophage were isolated by recombination in vivo. A HindIII restriction fragment of 8 kilobases (kb) previously shown to contain at least part of the rho gene (Gray et al. 1981) was cloned into plasmid pMC81. This vector has transcription stop sites that present read-through expression of cloned genes from either direction, and cloning sites upstream of the lacZ gene coding for beta-galactosidase. The position of the rho gene and flanking sequences required for its expression were further localized to a region of approximately 2 kb by introducing deletions using restriction enzyme treatment of these plasmids. A promoter in the rho region was found to direct beta-galactosidase synthesis in these plasmid derivatives. Derivatives of lambda ilv-rho phage were isolated in vivo by pyrophosphate chelation selection for phage with reduced genome size. Restriction enzyme analysis of twelve of these derivatives revealed an unexpected bias towards phage recombinants as opposed to simple internal deletions.

Autoregulation of the rho gene of Escherichia coli K-12.

It has previously been proposed, based on indirect evidence, that the Rho protein may control the expression of the rho gene. Using an in vitro system for the transcription and translation of the rho gene cloned into plasmid pBR322, we tested this hypothesis directly by monitoring the effect in vitro of excess or limiting Rho protein. The addition of purified Rho protein suppresses Rho synthesis in vitro. The addition of antibody to Rho specifically stimulates Rho synthesis in vitro. The stimulation of Rho factor synthesis by antibody to Rho is reversed by Rho protein. Rho factor purified from a strain with a mutationally altered rho gene (rho-115) does not suppress Rho synthesis in vitro. These results provide convincing evidence that the rho gene is subject to autoregulation.

Role of cryptic genes in microbial evolution.

Cryptic genes are phenotypically silent DNA sequences, not normally expressed during the life cycle of an individual. They may, however, be activated in a few individuals of a large population by mutation, recombination, insertion elements, or other genetic mechanisms. A consideration of the microbial literature concerning biochemical evolution, physiology, and taxonomy provides the basis for a hypothesis of microbial adaptation and evolution by mutational activation of cryptic genes. Evidence is presented, and a mathematical model is derived, indicating that powerful and biologically important mechanisms exist to prevent the loss of cryptic genes. We propose that cryptic genes persist as a vital element of the genetic repertoire, ready for recall by mutational activation in future generations. Cryptic genes provide a versatile endogenous genetic reservoir that enhances the adaptive potential of a species by a mechanism that is independent of genetic exchange.

Cloning of the ilvA538 gene coding for feedback-hypersensitive threonine deaminase from Escherichia coli K-12.

A variety of experimental results implicate the ilvA gene product, threonine deaminase, as an autoregulatory protein that affects the expression of its own gene and those coding for some related proteins. Some of the most direct evidence comes from the analysis of mutations in the ilvA gene with pleiotropic genetic regulatory effects. The most extensively documented mutation, ilvA538, lowers the expression of and abolishes repression control of the ilvGEDA transcription unit. A pleiotropic effect of the ilvA538 mutation, which may be either incidental or mechanistically related to the loss of repression control, renders threonine deaminase feedback hypersensitive to the inhibition of catalytic activity by the pathway end product, isoleucine. We transferred this mutation to lambda dilv phage and pBR322 derivatives. Direct enzyme assay of the plasmid- and phage-coded ilvA538 gene product in delta ilv hosts confirmed the feedback hypersensitivity of the enzyme product. In conjunction with the ilvG671 (phenotype, ILvG+ Valr; previously designated ilvO671) allele located in cis, high levels of the plasmid and lambda dilv phage-coded mutant enzyme suitable for protein purification were observed. Deletion mapping experiments with lambda dilv phage confirmed that the ilvA538 mutation, and not mutations promoter proximal to ilvD (transcription is from ilvG to ilvA), confer a loss of repression control. These genetic mapping studies indicate, however, that an additional mutation(s) may be present that contributes, at least in part, to the reduced enzyme levels in strains with the ilvA538 mutation.