Faculty Bibliography

Translational stop codons were introduced at various locations in the protein-coding regions of the monocistronic bla and ompA gene transcripts of Escherichia coli, and the decay characteristics of the upstream and downstream mRNA segments were analyzed. Premature termination of translation at codon position 26 reduced the stability of both the translated and ribosome-free segments of bla mRNA, whereas release of ribosomes just 30 codons further downstream resulted in normal stability for both segments. Normal stability of an untranslated bla gene mRNA segment required its linkage to a ribosome-bound segment of bla gene mRNA. These findings indicate that depriving an mRNA segment of ribosomes does not necessarily render it more susceptible to degradation. However, premature termination of translation at a location that allows ribosomes to traverse only a short segment of bla mRNA can lead to destabilization of the entire transcript

To map the structural features responsible for the 5-fold difference in stability of the E. coli ompA and bla gene transcripts, we have constructed gene fusions that encode chimeric ompA/bla transcripts and a deletion that eliminates a large internal segment of bla mRNA. Shortening of bla transcripts by internal deletion or replacement of the 3' end with the corresponding segment of the ompA transcript had little effect on bla mRNA stability. However, fusion of a 5'-terminal 147 nucleotide segment of the ompA message 5' to full-length or truncated bla transcripts increased the half-life of the bla segments 3- to 5-fold. These and other findings indicate that E. coli transcripts contain discrete structural determinants of stability and instability that can influence the decay rate of linked mRNA segments derived from other genes

To probe the nature of the interconversion of the two unliganded forms of proline racemase, a number of experiments have been performed under oversaturating conditions where the rate of the enzymic reaction is mainly limited by the rate of this interconversion. Competitive deuterium washout experiments, where an equimolar mixture of D- and L-proline (in which some or all of one enantiomer is specifically deuterated at the 2-position) is allowed to reach chemical and isotopic equilibrium mediated by the enzyme, have been followed in four ways. The size and the rate of achievement of the maximum perturbation in the optical rotation have been measured, the deuterium content of the substrate at this maximum has been determined, and the final approach to equilibrium after the perturbation maximum has been followed. Further, the enzyme-catalyzed rate of tritium loss from [2-3H]proline has been established. Finally, it has been shown that the enzyme interconversion reaction is catalyzed by several buffers (such as ammonium, hydrazinium, and hydrogen sulfide). These data are discussed in terms of Marcus' theory, which allows a rather detailed picture of the mechanism of free enzyme interconversion to be drawn. This process nicely parallels the mechanism of the enzyme-catalyzed interconversion of the proline enantiomers, and it is evident that substrate racemization (with the concomitant switch of the enzyme-bound protons) is mirrored by the water-mediated switch of the enzyme-bound protons that effects the interconversion of the free enzyme forms. The results favor a stepwise reaction for the interconversion of the free enzyme forms in which a proton is abstracted from a bound water molecule to give a reaction intermediate having a hydroxide ion bound to the diprotonated form of the enzyme

The fractionation factors of protons bound to the essential catalytic groups in proline racemase have been determined by comparison of the time courses of two competitive deuterium washout experiments. The rate of achievement of the maximum perturbation in the optical rotation has been measured in the oversaturated region (that is, at high substrate concentrations) under two conditions: in the first, we start with an equimolar mixture of deuterated substrate S' and of unlabeled product P; in the second, we again start with equal concentrations of substrate and product, but the concentration of the deuterated material S' is less than 20% that of S. The different concentrations of deuterated substrate produce different levels of deuteration of the enzyme's catalytic groups, the kinetic consequence of which allow the fractionation factors of these enzymic groups to be determined. The observed values for the fractionation factors of the enzyme's groups of 0.55 +/- 0.1 are only consistent with these groups' being thiols. This conclusion is supported by results of measurements of the solvent isotope effect determined in the unsaturated regime. These findings confirm the earlier suggestion of Abeles and his group that two cysteine residues mediate the catalysis of proline racemization by this enzyme

To test whether a reaction involving the making and/or breaking of two bonds at two sites is concerted (and proceeds through a single transition state) or is stepwise (and involves a reaction intermediate in which only one bond has been made or broken), we have measured the isotopic fractionation at one site as a function of isotopic substitution at the other site. In the case of proline racemase, the discrimination against solvent deuterium in the product when the reaction is run in mixed H2O-D2O is measured for the reaction both of [2-1H]proline and of [2-2H]proline. The isotopic fractionation at the solvent site may in principle be smaller, the same, or larger, when the 2H-labeled substrate is used rather than the 1H substrate, and--depending upon the nature of the catalyzing groups--this information indicates whether the reaction is stepwise, or concerted, or whether an isotopically insensitive transition state is partially rate determining. Experimentally, we have found that the discrimination against solvent deuterium in the product L-proline is the same, whether D-[2-1H]proline or D-[2-2H]proline is the substrate. This result requires that the substrate and product 'on-off' steps are faster than the racemization step and that the racemization reaction proceeds either in a concerted manner or in a stepwise fashion involving enzyme catalytic groups (e.g., thiols) having ground-state fractionation factors around 0.5

The isotope effects for the interconversion of L-proline and D-proline, catalyzed by proline racemase, have been determined in the saturated region with both [2-2H]proline and [2-3H]proline. The deuterium fractionation factors for each of the protons in flight have been obtained from two kinds of experiment: by measuring the rate of racemization of one [2-2H]proline enantiomer as it racemizes into an equilibrated pool of unlabeled proline and by measuring the deuterium content of a proline sample at the optical rotation maximum that occurs when an equimolar mixture of one deuterium-labeled enantiomer and the other unlabeled enantiomer runs to equilibrium. The tritium fractionation factors for each of the protons in flight have been determined from measurements of the rate of loss of tritium to the solvent as one [2-3H]proline enantiomer runs to equilibrium. Good agreement is found among the fractionation factors determined by each method. The deuterium fractionation factors for the two protons are not identical: that for the proton derived from L-proline is 0.375 and that for the proton derived from D-proline is 0.44. This difference has been confirmed by a double-competition experiment in which the optical rotation of a mixture of DL-[2-2H]proline and unlabeled DL-proline is followed with time. The rotation (initially zero) passes through a maximum, from which the ratio of the two fractionation factors (0.86) is obtained. These data, coupled with the equilibrium fractionation factor for the 2-position of proline (which has been determined to be 1.17), provide the transition-state factors for each of the in-flight protons, and delineate the nature of the transition state(s) for the enzyme-catalyzed racemization

We report that the light-harvesting and reaction center genes in the rxcA locus of R. capsulata are contained within a single operon and that their differential expression results predominantly from marked segmental differences in stability within the polycistronic rxcA transcript. The 3' portion of this transcript is rapidly degraded to give rise to either of two slowly decaying mRNA remnants, both of which encode only the light-harvesting polypeptides. The greater stability of these remnants accounts for nearly all of the difference between the concentrations of the light-harvesting and reaction center proteins. The unstable 3' portion of the transcript is delimited by two alternative stem-and-loop structures, which apparently act as barriers to 3' exoribonucleases and thereby protect the upstream RNA segment. When a DNA fragment containing the rxcA locus was fused to a plasmid promoter and transcribed in E. coli, the long precursor transcript was processed to two short messages of greater stability, as in R. capsulata

The rate of production of bacterial gene products is known to vary with the rate of cell growth, the concentrations of many cellular proteins are altered during times of decreased growth rate. In addition, proteins whose in vivo levels show no significant alterations with changes in cell doubling time must be synthesized at rates that vary in direct proportion to the growth rate of the cell. In certain instances, growth-rate dependent gene regulation has been shown to occur at the transcriptional or translational level. Another potentially important element in the regulation of gene expression is the stability of messenger RNA. We report here the effect of bacterial growth rate on the half lives of four different monocistronic Escherichia coli mRNA species. The stabilities of two of these species, the transcripts of the ompA and cat genes, exhibited a marked dependence on cell growth rate, whereas the half lives of the transcripts of the lpp and bla genes are constant over a broad range of cell doubling times. Our results indicate that E. coli can alter the rate of synthesis of certain proteins by modulating mRNA stability in response to changes in the rate of cell growth

An assay was developed to investigate the fate of specific segments of beta-lactamase (bla) and ompA gene transcripts in Escherichia coli. DNA probes cloned in bacteriophage M13 were treated with an endonuclease capable of cleaving single-stranded DNA, the fragments produced were annealed with total cellular RNA, and the resulting RNA . DNA hybrids were subjected to S1 nuclease treatment and gel fractionation. By using this assay, direct evidence was obtained for 3'-to-5' directionality in the decay of the long-lived mRNA encoded by the ompA gene, and no preferential stability was observed for translated versus untranslated mRNA segments. In the case of bla mRNA, initial cleavage of the full-length transcript was rate limiting, and no decay intermediates were detected. No difference in degradation rate was seen for bla transcripts having variant 3' or 5' termini

The infrared spectrum of the complex of D-fructose 1,6-bisphosphate bound to yeast aldolase displays three spectral features between 1700 and 1800 cm-1. One of these (at 1730 cm-1) corresponds to the carbonyl group of enzyme-bound D-fructose 1,6-bisphosphate and/or dihydroxyacetone phosphate. The frequency of this band, which is unaffected by the removal of the intrinsic zinc ion from the enzyme, demonstrates that this carbonyl group is not significantly polarized when the substrate binds to the enzyme. In contrast, the spectral band assigned to the carbonyl group of enzyme-bound D-glyceraldehyde 3-phosphate (at 1706 cm-1) appears at a frequency 24 cm-1 lower than when this substrate is in aqueous solution. This shift indicates considerable polarization of the carbonyl group when D-glyceraldehyde 3-phosphate is bound at the active site. The third spectral feature (at 1748 cm-1), which is observed only in the presence of potassium ion, probably corresponds to an enzymic carboxyl group in a nonpolar environment