The absence of branched-chain amino acid and growth rate control at the internal ilvEp promoter of the ilvGMEDA operon.

The question of whether the promoter ilvEp, located in the coding region of ilvM, the second structural gene in the ilvGMEDA operon, is subject to either amino acid- or growth rate-mediated regulation is examined. The experiments described here were performed with ilvEp-cat and ilvEp-lac fusions carried as single copies on the chromosome. The activity of the ilvEp promoter was found to respond neither to the availability of branched-chain amino acids nor to a wide range of growth rates between 35 to 390 min. In the absence of any known role for the products of the ilvGMEDA operon when repressing levels of branched-chain amino acids are present, there appears to be only a gratuitous role for the transcription at ilvEp.

Mutations replacing the leucine codons or altering the length of the amino acid-coding portion of the ilvGMEDA leader region of Escherichia coli.

The specificity of regulation by attenuation of the ilvGMEDA operon of Escherichia coli was examined by making alterations in the peptide-coding portion of the leader region. The effects of the alterations on attenuation control were monitored by operon fusions with the lacZ or cat gene. Substitution of the tandem leucine codons with arginine codons did not result in arginine control of attenuation even though the altered leader transcripts contained three consecutive arginine codons. Substitution of the single leucine codon with a proline codon at position 10 of the putative peptide, which had been shown to be important in the regulation of the Serratia marcescens ilv operon, did not result in control of attenuation by proline. Since the formation of neither proline nor arginine biosynthetic enzymes is regulated by attenuation control, the effect of tandem phenylalanine codons in place of the tandem leucine codons was examined and found not to result in control by phenylalanine supply. The latter failure may have been due to a configuration in the secondary structure of the protector stem of the leader transcript different from that of the wild-type transcript. The results of the study favored the idea that the lead ribosome does not initiate translation of the leader transcript until after the RNA polymerase has reached the pause site (117 bases into the leader region).

Specificity of attenuation control in the ilvGMEDA operon of Escherichia coli K-12.

Three different approaches were used to examine the regulatory effects of the amino acids specified by the peptide-coding region of the leader transcript of the ilvGMEDA operon of Escherichia coli K-12. Gene expression was examined in strains carrying an ilvGMED'-lac operon fusion. In one approach, auxotrophic derivatives were starved of single amino acids for brief periods, and the burst of beta-galactosidase synthesis upon adding the missing amino acid was determined. Auxotrophic derivatives were also grown for brief periods with a limited supply of one amino acid (derepression experiments). Finally, prototrophic strains were grown in minimal medium supplemented with single and multiple supplements of the chosen amino acids. Although codons for arginine, serine, and proline are interspersed among the codons for the three branched-chain (regulatory) amino acids, they appeared to have no effect when added in excess to prototrophs or when supplied in restricted amounts to auxotrophs. Deletions removing the terminator stem from the leader removed all ilv-specific control, indicating that the attenuation mechanism is the sole mechanism for ilv-specific control.

Translational coupling between the ilvD and ilvA genes of Escherichia coli.

The hypothesis that translation of the ilvD and ilvA genes of Escherichia coli may be linked has been examined in strains in which lacZ-ilvD protein fusions are translated in all three reading frames with respect to ilvD. In these strains, the nucleotide sequence was altered to obtain premature termination of ilvD translation, and in one strain translation termination of ilvD DNA occurred two bases downstream of the ilvA initiation codon. In the wild-type strain, the ilvD translation termination site was located two bases upstream of the ilvA start codon. In each of the mutant strains, expression of ilvA, as determined by the level of threonine deaminase activity, was strikingly lower than in the wild-type strain. The data suggest that expression of ilvD and ilvA is translationally coupled. By inserting a promoterless cat gene downstream of ilvA, it was shown that the differences in enzyme activity were not the result of differences in the amount of ilvA mRNA produced.

Role of codon choice in the leader region of the ilvGMEDA operon of Serratia marcescens.

Leucine participates in multivalent repression of the Serratia marcescens ilvGMEDA operon by attenuation (J.-H. Hsu, E. Harms, and H.E. Umbarger, J. Bacteriol. 164:217-222, 1985), although there is only one single leucine codon that could be involved in this type of control. This leucine codon is the rarely used CUA. The contribution of this leucine codon to the control of transcription by attenuation was examined by replacing it with the commonly used leucine codon CUG and with a nonregulatory proline codon, CCG. These changes left intact the proposed secondary structure of the leader. The effects of the codon changes were assessed by placing the mutant leader regions upstream of the ilvGME structural genes or the cat gene and measuring acetohydroxy acid synthase II, transaminase B, or chloramphenicol acetyltransferase activities in cells grown under limiting and repressing conditions. The presence of the common leucine codon in place of the rare leucine codon reduced derepression by about 70%. Eliminating the leucine codon by converting it to proline abolished leucine control. Furthermore, a possible context effect of the adjacent upstream serine codon on leucine control was examined by changing it into a glycine codon.

Effects of deletion and insertion mutations in the ilvM gene of Escherichia coli.

A plasmid was constructed that carried the ilvG and ilvM genes and the associated promoter and leader regions derived from the K-12 strain of Escherichia coli. The ilvG gene contained a + 1 frameshift mutation that enabled the plasmid to specify acetohydroxyacid synthase II. The plasmid was modified by deletions in the terminus of and within the ilvM gene and by insertions into the ilvM gene. The effects of these modifications on the phenotypes of the plasmids were examined in a host strain that lacked all three isozymes of acetohydroxyacid synthase. Most of the ilvM mutant plasmids so obtained permitted growth of the host strain in the absence of isoleucine but not in the absence of valine. Growth in the presence of valine, however, was very slow. No significant acetohydroxyacid synthase activity could be detected even when the cells were grown in a valine-supplemented minimal medium. It thus appears that, at most, only a very low level of acetohydroxyacid synthase activity occurred with ilvG in the absence of ilvM and that low activity was more effective for acetohydroxy butyrate formation than for acetolactate formation. The ilvM gene product could be formed under the control of the lac promoter in the presence of a plasmid that carried an in-frame gene fusion between lacZ and the downstream portion of ilvG. Extracts from the host strain that contained such an IlvG(-)-IlvM+ plasmid could be combined with extracts from cells that contained one of the IlvG+-IlvM- plasmids to yield acetohydroxyacid synthase activity. Thus, the ilvM and ilvG genes could be expressed independently of each other.

Comparison of the regulatory regions of ilvGEDA operons from several enteric organisms.

The nucleotide sequence preceding the ilvGEDA operon has been examined and compared in five enteric organisms. The sequence in Escherichia coli B was identical to the earlier-described strain K-12 sequence. The sequences of Salmonella typhimurium and Klebsiella aerogenes were remarkably similar to that of E. coli and identical in that part of the leader region that specified the putative 32-amino-acid peptide. Thus, identical secondary structures could be postulated for the leaders of all three organisms, and regulation of operon expression could be like that postulated earlier for E. coli. Different secondary structures had to be postulated for the leader transcripts of Edwardsiella tarda and Serratia marcescens. Control of attenuation of the operon in these organisms by the level of leucyl tRNA could be explained only if ribosome stalling occurred at a single leucine codon. In both organisms, that single leucine codon is the rarely used CUA rather than the CUG that is in E. coli, S. typhimurium, and K. aerogenes.

Leucine regulation of the ilvGEDA operon of Serratia marcescens by attenuation is modulated by a single leucine codon.

The effect of leucine limitation and of restricted leucine tRNA charging on the expression of the ilvGEDA operon of Serratia marcescens was examined. In this organism, the ilv leader region specifies a putative peptide containing only a single leucine codon that could be involved in leucine-mediated control by attenuation (E. Harms, J.-H. Hsu, C. S. Subrahmanyam, and H. E. Umbarger, J. Bacteriol. 164:207-216, 1985). A plasmid (pPU134) containing the DNA of the S. marcescens ilv control region and three of the associated structural genes was studied as a single chromosomal copy in an Escherichia coli strain auxotrophic for all three branched-chain amino acids. The S. marcescens ilv genes responded to a multivalent control similar to that found in other enteric organisms. Furthermore, the S. marcescens ilv genes were derepressed when the charging of leucine tRNA was restricted in a leuS derivative of E. coli that had been transformed with pPU134. It was concluded that ribosome stalling leading to deattenuation is not dependent on either tandem or a consecutive series of codons for the regulatory amino acid. However, the fact that the single leucine codon is a less frequently used codon (CUA) may be important. The procedure for obtaining the cloned ilv genes in single chromosomal copy exploited the dependence of ColE1 replicons on the polA gene. The cloning experiments also revealed a branched-chain amino acid-glutamate transaminase in S. marcescens that is different from transaminase B.

Isolation and analysis of two Escherichia coli K-12 ilv attenuator deletion mutants with high-level constitutive expression of an ilv-lac fusion operon.

A lysogenizing lambda phage, lambda dilv-lac11, was constructed to carry an ilvD-lac operon fusion. Expression from the phage of the ilvE and lacZ genes is controlled by an intact ilv control region also carried by this phage. Two spontaneous mutants of lambda dilv-lac11 that have high-level constitutive expression of the ilv-lac fusion operon were isolated by growth on a beta-chloroalanine selective medium. The mutants were shown by nucleotide sequence determination to contain large deletions (delta 2216, approximately 1.6 kilobases; delta 2219, approximately 1.9 kilobases), which in both cases remove the proposed ilv attenuator terminator. The rest of the ilv leader and promoter region DNA remains intact in these mutants. Deletion 2216 also removed part of the downstream ilvG gene, whereas delta 2219 extended through the entire ilvG gene into the ilvGE intercistronic region. A possible mechanism of deletion formation is discussed.

Exogenous Valine Reduces Conversion of Leucine to 3-Methyl-1-Butanol in Saccharomyces cerevisiae.

Mutant strains of the yeast Saccharomyces cerevisiae that require branched-chain amino acids must be supplemented with large concentrations (up to 10 mM) of these amino acids to satisfy their nutritional requirement. The utilization of one branched-chain amino acid, leucine, was examined in several leul strains of yeast grown aerobically in a glucose-ammonium salts minimal medium containing a limiting concentration (0.2 mM) of leucine. In this medium, the leucine requirement of the auxotrophic strains could be reduced by valine, another branched-chain amino acid. Increasing the valine concentration increased the cell yields of cultures and also reduced the levels of 3-methyl-1-butanol detected in the medium by gas chromatography. The concentration of 3-methyl-1-butanol was reduced from 122.0 to 48.9 muM when 5.0 mM valine was supplemented to limiting-leucine cultures. The amino acids isoleucine, threonine, norleucine, norvaline, alpha-amino-butyrate, alanine, and glycine also spared the leucine requirement of leucine auxotrophs, most likely because they resembled leucine and competed for its uptake. We propose that leucine analogs restrict the entry and degradation of leucine and thus reduce its conversion to 3-methyl-1-butanol, a major component of fusel oil.

Physical and genetic localization of ilv regulatory sites in lambda ilv bacteriophages.

A set of nine lambda dilv phages were used to transduce bacterial recipients containing point mutations or deletions in the ilv genes located at 84 min on the Escherichia coli K-12 chromosome. This genetic analysis indicated that two phages carry the entire ilvGEDAC cluster; others carry the complete ilvC gene and, in addition, bacterial DNA that extends to a termination point between ilvA and ilvC, within ilvD, within ilvE, or within ilvG. DNA extracted from the lambda dilv phages was digested with EcoRI, HindIII, KpnI, PstI, SalI, and SmaI. The restriction maps revealed that these phages were generated after insertion at four distinct insertion sites downstream (clockwise) of ilvC. The physical relationships between the various phages were further examined by electron microscopic heteroduplex analysis. The physical maps of the phages thus generated were straightforward and in complete accord with the genetic data. No evidence for genetic rearrangements of ilv DNA in the phage was obtained, thus validating conclusions based on the use of these phages in previous and ongoing research projects. Bacterial cells with deletions of the ilv genes were made lysogenic with lambda dilv phage to examine the regulation of ilv genes present in the phage. The results confirm previous studies showing that one site for control by repression and derepression is upstream (counterclockwise) of ilvG. It was shown, in addition, that the activities of dihydroxy acid dehydrase and threonine deaminase were increased when the prototrophic lysogens were grown with 20 mM leucine. Since this increase was exhibited even when the ilvG-linked control region was not carried by the lambda dilv phage, additional control sites must be located within the ilvEDA region of the ilvGEDA transcription unit.

Mutations in the ilvY gene of Escherichia coli K-12 that cause constitutive expression of ilvC.

A derivative of Escherichia coli K-12 bearing an ilvC-lac fusion has been studied. beta-Galactosidase formation in this strain is under the control of the ilvC promoter and is therefore induced by the acetohydroxy acids. Derivatives of this fusion strain were isolated that constitutively expressed beta-galactosidase. When an ilvC-containing episome was introduced into these strains, acetohydroxy acid isomeroreductase was also constitutively expressed. The lesions are trans dominant and lie in ilvY, the structural gene specifying a positive control element, v, needed for induction of the isomeroreductase. It was concluded from measurements of beta-galactosidase levels in various diploid strains that, although wild-type v requires inducer to act as a positive control element, it does not act as a repressor in the absence of inducer.

In vitro formation of beta-galactosidase with a template containing the lac genes fused to gene ilvD.

An in vitro coupled transcription-translation system was used to synthesize transaminase B and beta-galactosidase in the presence of a deoxyribonucleic acid template containing lac deoxyribonucleic acid under normal lac-specific control and in the presence of several deoxyribonucleic acid templates containing lac deoxyribonucleic acid fused to the ilvD gene. Time course experiments revealed that transcription of the lacZ gene from the fusion template required a longer time than did that initiated at the lac promoter. With a phage template containing an intact ilvE gene but lacking the normal ilv-specific promoter, synthesis of ilvE message was completed before synthesis of lacZ message. A phage template that contained the normal ilv-specific promoter but from which part of ilvE had been deleted also allowed formation of beta-galactosidase. Three plasmids containing the ilv-lac fusion were also used as templates. Two plasmids that contained both an intact ilvE gene and the normal ilv-specific promoter required longer times for lacZ transcription but were more efficient templates than was a plasmid in which the ilv-lac fusion, the ilvE gene, and the contiguous non-specific ilvE promoter were inverted with respect to the normal ilv-specific promoter. beta-Galactosidase synthesis was stimulated by guanosine 3'-pyrophosphate-5'-pyrophosphate with all templates tested except that in which the ilv-lac fusion had been inverted. Presumptive evidence was obtained for the generation of a limiting isoleucine signal by incorporating inhibitors of isoleucyl transfer ribonucleic acid synthetase into the coupled transcription-translation system.

Regulation of ilvEDA expression occurs upstream of ilvG in Escherichia coli: additional evidence for an ilvGEDA operon.

A low-copy-number plasmid was prepared that contained the entire ilv gene cluster of Escherichia coli. The introduction of an ilvO mutation allowed the ilvG gene of the plasmid to be expressed and imparted valine resistance to strains carrying it. Insertion of Tn10 into the ilvG gene of the plasmid resulted in a strong polar effect on ilv genes E, D, and A. Replacement of a region of ilv deoxyribonucleic acid between two KpnI sites on the high-copy-number plasmid carrying the entire ilv gene cluster with a KpnI fragment carrying an ilv-lac fusion but not extending into the ilv-specific control region resulted in a plasmid expressing the lacZ gene under ilv control when the fusion had been inserted in its normal orientation but not when it had been inserted in the opposite orientation. These experiments indicate that ilv-specific control over ilvE, ilvD, and ilvA expression is dependent on these genes being continguous with deoxyribonucleic acid that lies upstream of ilvG. The results also add further support to the concept of an ilvGEDA operon in E. coli.

Physical location of the ilvO determinant in Escherichia coli K-12 deoxyribonucleic acid.

A plasmid carrying the 4,6-kilobase (kb) HindIII-derived fragment from an ilvO mutant derivative of lambda h80dilv imparted a valine-resistant phenotype on strains it carried. This fragment carries a small amount of the promoter-proximal end of ilvE, the ilvO determinant, and apparently the entire ilvG gene, which specifies the valine-insensitive acetohydroxy acid synthase. Comparable deoxyribonucleic acid (DNA) from the original lambda h80dilv did not carry the valine resistance marker. The valine-resistant phenotype was always correlated with the formation of the resistant enzymes. The ilvO determinant was shown to be carried within an approximately 600-based-pair region lying between the SalI and KpnI sites on the HindIII fragment and perhaps within the ilvG gene itself. Ribonucleic acid that hybridizes with the DNA corresponding to the ilvG gene is formed in wild-type K-12 cells. This fact, coupled with the fact that ilvG is transcribed from the same DNA strand as the ilvE, D, and A genes, led to the idea that transcription is normally initiated upstream from ilvG in both wild-type and ilvO strains. In wild-type strains either the formation or the translation of the transcript would be terminated with the ilvG gene, thus preventing expression of that gene.

Location of the multivalent control site for the ilvEDA operon of Escherichia coli.

A strain of Escherichia coli K-12 containing a deletion extending from early in the ilvE gene toward the ilvG gene was shown to exhibit a higher expression of the downstream genes, ilvD and ilvA, than did an ilv+ strain. The elevated expression was under apparently normal ilv-specific control, however. The deletion was transferred to the ilv region of lamba h80dilv and shown by restriction endonuclease and heteroduplex analysis to extend through the deoxyribonucleic acid (DNA) shown, in the preceding paper (C. S. Subrahmanyam, G. M. McCorkle, and H. E. Umbarget, J. Bacteriol 142:547--555, 1980), to contain the ilvO determinant. The deletion was also transferred to an ilv-lac fusion strain and shown to cause an increase in beta-galactosidase formation while allowing retention of ilv-specific control. Transducing phages excised from these fusion strains with and without the ilvO determinant were compared. The phage carrying the ilvO+ determinant contained ilv DNA extending only into but not through the ilvG gene. It did not exhibit an ilv-specific control of beta-galactosidase formation. The phage carrying the deletion of ilvO but containing ilv DNA extending beyond the ilvG gene exhibited ilv-specific control of beta-galactosidase formation. It was concluded that the multivalently controlled ilv-specific promoter affecting ilv operon expression lies upstream from ilvG and that the ilvO region in the wild-type K-12 strain is a region of polarity preventing ilvG expression and reducing ilvEDA expression.

Nucleotide sequence of ilvGEDA operon attenuator region of Escherichia coli.

The nucleotide sequence of the DNA thought to contain the control region for the ilvGEDA operon in Escherichia coli has been determined by the Maxam-Gilbert procedure. The sequence includes a region that, upon transcription, would yield a leader transcript specifying a peptide 32 residues long. This putative peptide would contain four leucine, five isoleucine, and six valine residues. A model is proposed that correlates the multivalent control of the ilvGEDA operon with the extent to which this leader transcript is translated. In vitro transcription experiments yielded a transcript of about 183 nucelotides, compatible with the predictions of the model.

Physical characterization of ilv-lac fusions.

Electron microscopic heteroduplex analysis and comparative restriction digests have been used to characterize lambda p123(209), the complementary pair of phages used in the Casadaban technique of gene fusion. Derivatives of lambda 1(209) constructed to carry fusions of the lac genes to the control regions of the ilvC and ilvEDA operons were also analyzed. These physical maps have provided confirmation of the genetic models for these constructions and physical specifications important in interpreting the behavior of these ilv-lac fusions.

ilvU, a locus in Escherichia coli affecting the derepression of isoleucyl-tRNA synthetase and the RPC-5 chromatographic profiles of tRNAIle and tRNAVal.

A mutation in the ilvU locus of Escherichia coli has led to a complex phenotype that included resistance to thiaisoleucine, a loss of derepressibility of isoleucyl tRNA synthetase, and an alteration of the RPC-5 chromatographic profile of the branched-chain aminoacyl-tRNA's. The alterations were manifest in an increase in the amount of Species 2 of both tRNAIle and tRNAVal at the expense of Species 1. A similar alteration, but independent of (and additive to) that caused by the ilvU mutation, was observed upon limitation of either isoleucine or valine. The shift in profile caused by limitation was also independent of the reduced growth rate or the derepression of the isoleucine and valine biosynthetic enzymes that also result from limitation. During chloramphenicol treatment nearly all tRNAIle and tRNAVal formed appears as species 2. Upon recovery from chloramphenicol, Species 2 of both acceptors are converted to Species 1. It is proposed that the ilvU product not only allows derepression of isoleucyl-tRNA synthetase but also retards the conversion of tRNA2Ile to tRNA1Ile and that of tRNA2Val to tRNA1Val. The mutated ilvU loci abolish the derepression and are more efficient in retarding the conversion.