Number Of RNA Polymerase Molecules Research Articles

A counting-strategy together with a spatial structured model describes RNA polymerase and ribosome availability in Escherichia coli

The allocation of resources during bacterial growth is strongly related to the availability of ribosomes and RNA polymerase molecules. Here, coarse-grained models offer a promising start due to their simple structure and the limited number of kinetic parameters. Based on published data sets for proteome and mRNA data in Escherichia coli, and together with mass balance equations describing gene expression, we are able to calculate the number of active molecules (that is, the number of ribosomes that are currently translating nascent and mature mRNA, as well as the number of RNA polymerase molecules on the DNA). This information is a prerequisite for meaningful coarse-grained models. In our approach, the cellular compartment is structured into a cytosolic region and a nucleoid region, and the processes of transcription and translation are assigned accordingly. The theoretical study reveals a quadratic relationship between the number of active ribosomes and the growth rate μ. While the overall available number of ribosomes follows the linear “bacterial growth law”, the approach allows us to determine the growth limit for the chosen experimental environment (minimal medium, only one C source). The new approach is in good agreement with published experimental data, and, with a simple rule of thumb can be applied to other cellular systems.

The activation of amplified ribosomal RNA genes in the oocytes of Xenopus laevis: an electron microscope analysis.

Mitchell, E. L. D. and Hill, R. S. 1987. The activation of amplified rihosomal genes in the oocytes of Xenopus laevis: an electron microscope analysis. —Hereditas 107 219–227. Lund, Sweden. ISSN 0018–0661. Received February 16, 1987 The Miller spreading technique has been used to study the ultrastructure of chromatin from the inactive, amplified ribosomal RNA genes found in the previtellogenic oocytes of Xenopus laevis and to see how this inactive structure is modified during the transcriptional activation of the genes in early vitellogenesis. Inactive, nucleolar chromatin fibres from 300 pm oocytes are packaged into nucleosomal and supranucleosomal structures of 20—-22 nm in diameter. However, where visible, the chromatin axis in active ribosomal genes has a smooth appearance and a diameter of 24 nm. By contrast, the axis of transcription units from lamphrush chromosomes, found in the same preparations, has a distinctively nucleosomal-type of organization. During transcriptional activation, in oocytes of 350–400 pm in diameter, the morphological appearance of the newly-activated genes varies considerably with respect to the length of the gene, the number of RNA polymerase molecules per gene and the amount of RNA transcript associated with the polymerases. However, in oocytes of at least 450 pm, the genes in most, but not all nucleoli, have the classical “Christmas Tree” appearance, consisting of densely packed RNA polymerases and transcripts which show a distinctive gradient in their length from the beginning to the end of the gene.

Transcriptional Polarity in rRNA Operons of Escherichia coli nusA and nusB Mutant Strains

Synthesis of ribosomes in Escherichia coli requires an antitermination system that modifies RNA polymerase to achieve efficient transcription of the genes specifying 16S, 23S, and 5S rRNA. This modification requires nucleotide signals in the RNA and specific transcription factors, such as NusA and NusB. Transcription of rrn operons in strains lacking the ability to produce either NusA or NusB was examined by electron microscopy. The distribution and numbers of RNA polymerase molecules on rrn operons were determined for each mutant. Compared to the wild type, the 16S gene in the nusB mutant strain had an equivalent number of RNA polymerase molecules, but the number of RNA polymerase molecules was reduced 1.4-fold for the nusA mutant. For both mutant strains, there were twofold-fewer RNA polymerase molecules on the 23S RNA gene than for the wild type. Overall, the mutant strains each had 1.6-fold-fewer RNA polymerase molecules on their rrn operons than did the wild type. To determine if decreased transcription of the 23S gene observed by electron microscopy also affected the 30S/50S ribosomal subunit ratio, ribosome profiles were examined by sucrose gradient analysis. The 30S/50S ratio increased 2.5- to 3-fold for the nus mutant strains over that for wild-type cells. Thus, strains carrying either a nusA mutation or a nusB mutation have defects in transcription of 23S rRNA.

The effect of prostaglandin E1 on in vitro transcription of sperm chromatin, isolated from patients with azoospermia, teratospermia and chronic prostatitis.

We have investigated the influence of Prostaglandin E1 on the in vitro transcription of chromatin, isolated from spermatozoa of patients suffering from different pathologies, leading to infertility, namely, azoospermia, teratospermia and chronic prostatitis. Our studies indicate that prostaglandin E1 has a stimulatory effect both on in vitro transcription, on the number of RNA polymerase molecules and the polyribonucleotide elongation rates as compared to sperm chromatin from healthy patients. The results on the incorporation of alpha-32P-ATP in to RNA in the presence and absence of Prostaglandin E1 correlate well with the data on the number of actively transcribing RNA polymerase molecules and the rate of RNA elongation, which might be due to low levels of prostaglandin E1 in human semen.

Ultrastructural analysis of Balbiani ring genes of Chironomus pallidivittatus in different states of Balbiani ring activity

Patterns of gene activity on Balbiani ring (BR) transcription units (TUs) of polytene chromosome IV of Chironomus pallidivittatus larvae were ultrastructurally characterized by electron microscopy. In Miller chromatin speads, the 9 μm long TUs of Balbiani ring 1 and 2 (BR1 and BR2), respectively, were unambiguously identified on the basis of the size and structure of their stalked ribonucleoprotein (RNP) granules. Because of the mild spreading procedure the morphology of these 50 nm BR RNP particles remained quite reminiscent of their in situ configuration. BR TUs were found as clusters of sister copies, looking almost homogeneous in structure and organization, a number of them appearing as loops. TUs of normal and hyperactive BRs (the latter resulting from stimulation of the larvae with pilocarpine) were compared. The patterns of transcriptional activity of BR TUs were rather similar but not absolutely identical on all homologous sister chromatin strands. In normal as well as in hyperactive BRs, the great majority of TUs were highly transcriptionally active, i.e. densely packed with RNP fibres all along the deoxyribonucleoprotein (DNP) axis. Mean RNP fibre frequency of BR TUs of normal larvae was 30/μm while hyperactive BRs showed a fibre frequency of 31μm. The average number of RNP fibres per BR TU is in the range of 200 to 300. Assuming a transcription time of 20 min per BR mRNA molecule, the number of RNA polymerase molecules starting transcription of the 37 kb gene could be estimated to be 10 to 15 per minute. A number of BR genes were only partially active, exhibiting short, middle or long arrays of RNP fibres. Such arrays occurred in any segment of the genes showing normal RNP fibre density and a nucleosomal configuration in the inactive part of the gene. The variations in RNP fibre frequencies suggest that transcription can be modulated at the level of the individual BR gene. However, the large increase in the rate of BR mRNA synthesis upon pilocarpine stimulation may mainly be due to activation of hitherto inactive sister copies of the BR gene.

Transcriptional Regulation of Acetyl Coenzyme A Carboxylase Gene Expression by Tumor Necrosis Factor in 30A-5 Preadipocytes

Acetyl coenzyme A (acetyl-CoA) carboxylase activity, amount, and mRNA levels increase during the differentiation of 30A-5 preadipocytes to adipocytes. Tumor necrosis factor (TNF) completely prevents this differentiation, with concomitant inhibition of acetyl-CoA carboxylase mRNA accumulation. To investigate the mechanisms by which TNF prevents acetyl-CoA carboxylase mRNA accumulation, we determined the effect of TNF on the transcription rate of the carboxylase gene and the half-life of carboxylase mRNA. Nuclear runoff transcription assays revealed no differences in the number of RNA polymerase molecules actively engaged in transcription of the acetyl-CoA carboxylase gene in preadipocytes, adipocytes, TNF-treated preadipocytes, or at any time during the course of differentiation. However, changes in adipsin, glycerophosphate dehydrogenase, and actin mRNAs, whose levels are also differentiation dependent, can be accounted for in part by changes in the number of polymerase complexes on their respective genes. To determine whether TNF caused a decrease in the stability of carboxylase RNA transcripts, we measured the rate of decay of prelabeled acetyl-CoA carboxylase mRNA. Control and TNF-treated cells showed no difference between the apparent half-lives of acetyl-CoA carboxylase mRNAs (9 h). However, the rate of acetyl-CoA carboxylase mRNA synthesis in vivo was decreased three- to fourfold in the presence of TNF. These data demonstrate that TNF prevents accumulation of acetyl-CoA carboxylase mRNA during preadipocyte differentiation by decreasing the rate of acetyl-CoA carboxylase gene transcription. However, transcriptional control is not due to a change in the number of RNA polymerase complexes actively engaged in carboxylase transcript elongation which could be measured by a number runoff assay. Instead, transcriptional control may be related to the rate at which RNA polymerase traverses the acetyl-CoA carboxylase gene.

Transcriptional regulation of acetyl coenzyme A carboxylase gene expression by tumor necrosis factor in 30A-5 preadipocytes.

Acetyl coenzyme A (acetyl-CoA) carboxylase activity, amount, and mRNA levels increase during the differentiation of 30A-5 preadipocytes to adipocytes. Tumor necrosis factor (TNF) completely prevents this differentiation, with concomitant inhibition of acetyl-CoA carboxylase mRNA accumulation. To investigate the mechanisms by which TNF prevents acetyl-CoA carboxylase mRNA accumulation, we determined the effect of TNF on the transcription rate of the carboxylase gene and the half-life of carboxylase mRNA. Nuclear runoff transcription assays revealed no differences in the number of RNA polymerase molecules actively engaged in transcription of the acetyl-CoA carboxylase gene in preadipocytes, adipocytes, TNF-treated preadipocytes, or at any time during the course of differentiation. However, changes in adipsin, glycerophosphate dehydrogenase, and actin mRNAs, whose levels are also differentiation dependent, can be accounted for in part by changes in the number of polymerase complexes on their respective genes. To determine whether TNF caused a decrease in the stability of carboxylase RNA transcripts, we measured the rate of decay of prelabeled acetyl-CoA carboxylase mRNA. Control and TNF-treated cells showed no difference between the apparent half-lives of acetyl-CoA carboxylase mRNAs (9 h). However, the rate of acetyl-CoA carboxylase mRNA synthesis in vivo was decreased three- to fourfold in the presence of TNF. These data demonstrate that TNF prevents accumulation of acetyl-CoA carboxylase mRNA during preadipocyte differentiation by decreasing the rate of acetyl-CoA carboxylase gene transcription. However, transcriptional control is not due to a change in the number of RNA polymerase complexes actively engaged in carboxylase transcript elongation which could be measured by a number runoff assay. Instead, transcriptional control may be related to the rate at which RNA polymerase traverses the acetyl-CoA carboxylase gene.

Interspersed Homologous DNA of Autographa californica Nuclear Polyhedrosis Virus Enhances Delayed-Early Gene Expression

The five regions of homologous DNA which are interspersed in the genome of the baculovirus Autographa californica nuclear polyhedrosis virus increased the expression of a delayed-early gene of this virus. Although this activity was first observed as a 10-fold trans effect, the homologous region 5 (hr5) enhanced the expression of linked genes 1,000-fold. The hr5 enhancer also exhibited the other characteristics associated with viral enhancer elements, including orientation independence and the abilities to function at a distance from the linked promoter, to regulate heterologous promoters, and to increase the number of RNA polymerase molecules transcribing the linked genes. The expression of the immediate-early regulatory gene was not enhanced by cis-linked hr5, although the enhancer function may require the immediate-early regulatory gene product. The hr5 enhancer was relatively insensitive to competition by an excess of enhancer molecules. The nucleotide sequence of hr5 revealed two different conserved repeats separated by nonhomologous DNA. Deletion analysis of the hr5 enhancer indicated that a 30-base-pair inverted repeat was essential for enhancer function.

Transcription of Xenopus borealis rRNA genes in nuclei of Xenopus laevis oocytes

Recombinant plasmids that contained a single rRNA gene and spacer from Xenopus borealis were injected into the nuclei of Xenopus laevis oocytes. After spreading the contents of injected nuclei for electron microscopy, we found transcription units indicative of the accurate transcription of the borealis gene. Both the number of RNA polymerase molecules per transcription unit and the total number of active genes per nucleus were similar to those found for injected laevis genes. We also injected laevis oocytes with a plasmid that contained both a shortened laevis rRNA gene and a full-length borealis rRNA gene. Visualization of transcription on this plasmid showed that usually either a short or a long transcription unit was present on any one molecule. The morphologies of the short and long transcription units were consistent with their representing transcription of the laevis and borealis genes, respectively. Both types of transcription unit were present in roughly equal numbers. Since the laevis oocyte transcribed injected borealis rRNA genes as efficiently as it transcribed injected laevis rRNA genes, it stands in marked contrast to the situation in embryos of interspecific hybrids between laevis and borealis where transcription occurs only from the laevis rRNA genes. Possible explanations for the difference in transcriptional behavior of the borealis rRNA gene in hybrid embryos and laevis oocytes are discussed.

Molecular mechanisms of glucocorticoid hormone action

In the current studies, several aspects of glucocorticoid action have been investigated. The data provide further evidence for the notion that under physiological conditions the glucocorticoid-receptor interaction in the intact cell is driven primarily by hydrophobic interactions and that the initial steroid-receptor interaction rather than transport process or subsequent hormone-induced conformational changes in the receptor-steroid complex associated with activation or chromatin binding of the complexes is the limiting determinant for the overall steroid-binding reaction in the intact cell. Agonists require longer to bind to receptors than the antagonist progesterone and remain bound longer than the antagonist; this may reflect either agonist-induced conformational changes associated with the initial agonist-receptor reaction that are not induced by the antagonist or else a requirement for a more prolonged receptor residency time for agonist actions. The receptor is a single polypeptide chain of 87 000 molecular weight and a domain of it required for receptor action is probably on the carboxy-terminal portion of the molecule and is distinct from the steroid- and DNA-binding domains. The receptor-glucocorticoid complex binds to chromatin acceptor sites that involve DNA and other chromatin functions (probably proteins) and that are in excess over the receptor. These interactions lead to changes in chromatin structure that may reflect the primary receptor effect. The chromatin interactions can then lead to an increase in the number of RNA polymerase molecules engaged in transcribing a glucocorticoid-regulated gene (growth hormone). This can lead to an increase in pre-mRNA and consequently the mRNA product of the gene.

PpGpp regulates the binding of two RNA polymerase molecules to the tyrT promoter.

Bacterial promoters differ in the number of RNA polymerase molecules that bind to form a filterable polymerase-promoter complex. We show that two holoenzyme molecules interact with the tyrT promoter, probably as a dimer. This interaction is inhibited by ppGpp. By contrast a single holoenzyme monomer suffices for complex formation at the lacUV5 promoter. We propose that In vivo promoter selection by monomeric and dimeric forms of the enzyme could coordinate the synthesis of stable RNA with that of mRNA and could also account in part for the switch in transcriptional selectivity during the stringent response.

Control of ribosomal RNA synthesis in Escherichia coli

Nucleoids were isolated from Escherichia coli B/r cells in steady-stage growth at different rates. The number of RNA chains growing on each nucleoid was estimated from the amount and size of RNAs synthesized by endogenous RNA polymerase. This figure represents the number of RNA polymerase molecules that are functioning in transcription and thus serves to indicate the frequency of transcription in the cells. With an increase in the growth rate from 0.27 to 1.73 (h-1), (a) the number of total RNA chains per genome increases from about 252 to 838, (b) the number of ribosomal RNA (rRNA) chains per genome increases as a function of the second power of the growth rate from about 19 to 255, and (c) the number of non-rRNA chains per genome increases linearly from about 233 to 538.

Transcription and translation initiation frequencies of the Escherichia coli lac operon

We have calculated the number of RNA polymerase molecules transcribing the induced lac operon of Escherichia coli as well as the distance between ribosomes on the proximal z and on the distal a messages † † Abberviations used: message, mRNA coding for a single, functional polypeptide (a polycistronic mRNA carries several messages); βGase, β-galactosidase (EC 3.2.1.23) coded from z gene; GATase, galactoside acetyltransferase (EC 2.3.18) coded from the a gene. . These values were derived from: (a) rates of induced enzyme synthesis corrected by known turnover numbers to give numbers of enzyme monomers produced per cell per second; (b) the mass of each message derived either directly by hybridization of long-labeled RNA to specific DNA or from the rates of synthesis and decay obtained by hybridization of pulse-labeled RNA; the latter gives about 13 times better resolution; (c) conversion of message mass into a number of funtioning 3′ ends (producing finished polypeptides); size analyses indicate that the z message probably decays by a net directional degradation in the 5′ to 3′ direction. From this the fraction of completed molecules that are full-length and the total number with functional 3′ ends can be derived. (d) Rates of ribosome movement. Calculated values follow. (1) With a 3.3-second interval between transcription initiations, there are 38 molecules of RNA polymerase on z DNA per cell; with 1.7 copies of lac DNA, this gives 23 on each z cistron and five or six on the y and on the a cistrons. This corresponds to a spacing of about 135 nucleotides between polymerases, which is similar to that on the ribosomal cistrons which have only 1.7 seconds separating initiations. (2) There are 20 molecules of β-galactosidase monomer produced per cell per second from 1.1 × 10 −4 pg of z RNA. 38% of the message molecules are nascent and not producing enzyme. Half of the completed molecules and 70% of their mass are intact z messages to give 62 molecules of z message containing functional 3′ ends/cell. Ribosomes load onto these messages at 3.2-second intervals to give a ribosomal spacing of 110 nucleotides. An average of 40 ribosomes translate each z message with half of the message molecules translated by more than 28 ribosomes and half by less. (3) Only 2.4 molecules of galactoside acetyltransferase monomer are produced per cell per second from 1.8 × 10 −5 pg of a RNA. Only 20% of these molecules are nascent and 80% of the completed ones are intact; this gives 40 molecules of functioning a messages per cell. Ribosomes load to a message at 16-second intervals to give a spacing of 580 nucleotides. These results show that the frequencies of translation initiations can differ for different messages. The faster decay of a compared to z message is consistent with a model in which a ribosome can protect a vulnerable site near the start of a message from inactivation; messages that load less frequently decay faster. The combined net effect of these causally related processes could account for natural polarity (Zabin & Fowler, 1970), i.e. there is a fourfold lower production of enzyme from the a than from the z gene.

Transcription in vitro of DNA in isolated bacterial nucleoids

The in vitro transcription of DNA in isolated nucleoids from Escherichia coli was investigated. The purpose of the study was to determine the effects on DNA transcription of using as template a condensed bacterial chromosome. To compare the template capacity of the DNA in the condensed chromosome with that of the completely unfolded chromosome, very high RNA polymerase/DNA ratios were used during transcription so that the amount of DNA was limiting. The rate of total RNA synthesis was greater on the condensed chromosome. The RNA chain elongation rate and the average size of the RNA transcripts were similar on the condensed and unfolded templates; however, RNA chain initiation rates were greater on the condensed chromosome. The results show that the maximum number of RNA polymerase molecules which can simultaneously synthesize RNA is greater on the condensed chromosome. A preferential synthesis of ribosomal RNA was observed on both the condensed and unfolded chromosomes. Although the absolute rate of rRNA synthesis was greater on the condensed template, the rate of rRNA synthesis relative to the rate of total RNA synthesis was similar on both templates. The results suggest that the genes specifying rRNA do not differ significantly in their availability for transcription relative to the remainder of the genome in folded and unfolded chromosomes. Moderate concentrations of spermidine stabilized the condensed state of the folded chromosome during transcription and the average size of the isolated nucleoids, as estimated from sedimentation studies and direct microscopic visualization, remained unchanged during transcription.

Joint propagation of the and ' subunits of RNA polymerase in Escherichia coli.

The core enzyme of the DNA dependent RNA polymerase from E. coli consists of four subunits, 2α, 1β and 1β′ (ref. 1). The two larger subunits, β and β′, are synthesized in a coordinate manner, possibly on a polycistronic mRNA, in the sequence β before β′ in growing E. coli cells2. The bacteria also have a surplus of these two subunits and, presumably, of the core enzyme at various rates of growth, the number of RNA polymerase molecules being two to five times the number of RNA molecules synthesized by the enzyme3. Assuming a pool of dormant polymerase, the question arises whether the enzyme molecules exchange their subunits while traversing this pool between termination and reinitiation of RNA synthesis as has been described for the protein synthesizing system, the ribosome and its subunits4, or whether the enzyme, once assembled from its subunits, is stable in growing cells. The answer might shed some light on regulation of RNA polymerase activity in vivo.

Regulation of ribonucleic acid synthesis in Escherichia coli B-r: an analysis of a shift-up. 1. Ribosomal RNA chain growth rates.

Abstract The chain growth rate for ribosomal RNA was determined for Escherichia coli B r growing in succinate (μ = 0.69 doublings/h), glucose (μ = 1.36) and glucose/ amino acids (μ = 2.10) medium. With increasing bacterial growth rate the chain growth rate increases from 4400 to 6300 nucleotides/min. These values are almost twofold higher than the chain growth rate reported for messenger RNA; this implies that, following a nutritional shift-up, the transfer of a relatively small number of RNA polymerase molecules from unstable to stable RNA genes along with the increase in the stable RNA chain growth rate is sufficient to account for the abrupt increase in the net rate of RNA synthesis. Furthermore, our calculations indicate that the linear density of polymerase molecules on the ribosomal DNA template increases with the bacterial growth rate, such that in rapidly growing bacteria all ribosomal RNA genes (48 copies at μ = 3) are nearly saturated with RNA polymerase.

Biosynthesis of the β and β′ subunits of RNA polymerase in Escherichia coli

The amounts of the β and β′ subunits of the DNA-dependent RNA polymerase relative to the amount of total protein synthesized have been determined under a number of growth conditions in two strains of Escherichia coli. The results of these measurements have been expressed as the relative rate of synthesis of core RNA polymerase, α p, assuming the four constituent subunits (2α, 1β and 1β′) to be synthesized in equivalent amounts. This quantity, α p, was found not to vary greatly with the growth rate μ. For glucose-grown cells of E. coli B/r (μ = 1.5 doublings/h) α p = 1.4%, corresponding to about 7000 molecules of core RNA polymerase per cell. For slowgrowing cells the value obtained for α p is lower and for fast-growing cells somewhat 3 higher. The comparison of these values with the number of RNA polymerase molecules estimated to be actively engaged in RNA synthesis indicates that both slow- and fast-growing cells contain a surplus of RNA polymerase, if the catalytic unit is assumed to be the monomer of core RNA polymerase. In addition to the measurements of cells during balanced growth at various rates, α p has been determined during the transition from one growth rate to another and during synchronous growth. During a shift-up the rate of synthesis of polymerase follows closely the rate of total protein synthesis, α p being nearly constant for a period of twenty minutes after the shift. In a synchronously dividing culture of E. coli B/r, α p was seen to be fairly constant during two cycles of synchronous division. It appears that α p is rather insensitive to the effect of gene doubling during the cell cycle.

Two regulatory models involving variable growth periods of β-galactosidase

We have calculated the number of RNA polymerase molecules transcribing the induced lac operon of Escherichia coli as well as the distance between ribosomes on the proximal z and on the distal a messages†. These values were derived from: (a) rates of induced enzyme synthesis corrected by known turnover numbers to give numbers of enzyme monomers produced per cell per second; (b) the mass of each message derived either directly by hybridization of long-labeled RNA to specific DNA or from the rates of synthesis and decay obtained by hybridization of pulse-labeled RNA; the latter gives about 13 times better resolution; (c) conversion of message mass into a number of funtioning 3′ ends (producing finished polypeptides); size analyses indicate that the z message probably decays by a net directional degradation in the 5′ to 3′ direction. From this the fraction of completed molecules that are full-length and the total number with functional 3′ ends can be derived. (d) Rates of ribosome movement.Calculated values follow. (1) With a 3.3-second interval between transcription initiations, there are 38 molecules of RNA polymerase on z DNA per cell; with 1.7 copies of lac DNA, this gives 23 on each z cistron and five or six on the y and on the a cistrons. This corresponds to a spacing of about 135 nucleotides between polymerases, which is similar to that on the ribosomal cistrons which have only 1.7 seconds separating initiations. (2) There are 20 molecules of β-galactosidase monomer produced per cell per second from 1.1 × 10−4 pg of z RNA. 38% of the message molecules are nascent and not producing enzyme. Half of the completed molecules and 70% of their mass are intact z messages to give 62 molecules of z message containing functional 3′ ends/cell. Ribosomes load onto these messages at 3.2-second intervals to give a ribosomal spacing of 110 nucleotides. An average of 40 ribosomes translate each z message with half of the message molecules translated by more than 28 ribosomes and half by less. (3) Only 2.4 molecules of galactoside acetyltransferase monomer are produced per cell per second from 1.8 × 10−5 pg of a RNA. Only 20% of these molecules are nascent and 80% of the completed ones are intact; this gives 40 molecules of functioning a messages per cell. Ribosomes load to a message at 16-second intervals to give a spacing of 580 nucleotides.These results show that the frequencies of translation initiations can differ for different messages. The faster decay of a compared to z message is consistent with a model in which a ribosome can protect a vulnerable site near the start of a message from inactivation; messages that load less frequently decay faster. The combined net effect of these causally related processes could account for natural polarity (Zabin & Fowler, 1970), i.e. there is a fourfold lower production of enzyme from the a than from the z gene.