Increased transcription rates correlate with increased ... - CiteSeerX

1 downloads 70 Views 165KB Size Report
leuB and argH mRNA synthesis were determined using half-lives and concentrations, during ..... mRNA concentrations and decay rates in the presence of.
Microbiology (2004), 150, 1457–1466

DOI 10.1099/mic.0.26954-0

Increased transcription rates correlate with increased reversion rates in leuB and argH Escherichia coli auxotrophs Jacqueline M. Reimers, Karen H. Schmidt, Angelika Longacre,3 Dennis K. Reschke and Barbara E. Wright Division of Biological Sciences, University of Montana, Missoula, MT 59812, USA

Correspondence Barbara E. Wright [email protected]

Received 1 December 2003 Revised 15 January 2004 Accepted 20 January 2004

Escherichia coli auxotrophs of leuB and argH were examined to determine if higher rates of transcription in derepressed genes were correlated with increased reversion rates. Rates of leuB and argH mRNA synthesis were determined using half-lives and concentrations, during exponential growth and at several time points during 30 min of amino acid starvation. Changes in mRNA concentration were primarily due to increased mRNA synthesis and not to increased stability. Four strains of E. coli amino acid auxotrophs, isogenic except for relA and argR, were examined. In both the leuB and argH genes, rates of transcription and mutation were compared. In general, strains able to activate transcription with guanosine tetraphosphate (ppGpp) had higher rates of mRNA synthesis and mutation than those lacking ppGpp (relA2 mutants). argR knockout strains were constructed in relA+ and relA mutant strains, and rates of both argH reversion and mRNA synthesis were significantly higher in the argR knockouts than in the regulated strains. A statistically significant linear correlation between increased rates of transcription and mutation was found for data from both genes. In general, changes in mRNA half-lives were less than threefold, whereas changes in rates of mRNA synthesis were often two orders of magnitude. The results suggest that specific starvation conditions target the biosynthetic genes for derepression and increased rates of transcription and mutation.

INTRODUCTION Nutritional stress results in the derepression, activation and transcription of specific genes, and therefore may be a major cause of mutations. Transcription can lead to mutation by driving localized supercoiling and the accumulation of DNA secondary structures containing unpaired and/or mispaired bases known to be vulnerable to mutations (reviewed by Wright, 2000). Using a new computer algorithm, it has been possible to predict mutation frequencies in derepressed genes of Escherichia coli auxotrophs and in the human p53 cancer gene (Wright et al., 2002, 2003). In E. coli, amino acid biosynthetic operons are repressed in the presence of their end-product amino acid. However, when the cells are starved for a required amino acid, the biosynthetic operon (or regulon) is deattenuated (leucine operon) or derepressed (arginine regulon) and activated. During leucine starvation, for example, leucine-charged tRNAs become limiting, causing ribosomes to stall at one of the leucine control codons. This deattenuates the leu operon, allowing transcription to continue into the 3Present address: Department of Medicine/Section of Nephrology, University of Chicago, Chicago, IL 60637, USA.

0002-6954 G 2004 SGM

Printed in Great Britain

structural genes of the operon. The stalled ribosomes and accumulated uncharged tRNALeu also simultaneously trigger the stringent response and the synthesis of the alarmone guanosine tetraphosphate (ppGpp). In turn, ppGpp activates specific derepressed genes encoding enzymes required to overcome the starvation conditions. The stringent response also redirects the cell’s resources toward survival by decreasing the synthesis of stable RNA, nucleotides and other metabolites required for cell replication. During the stringent response, ppGpp synthesis is catalysed by the relA gene product (ppGpp synthase I) and ppGpp begins accumulating immediately following starvation for any amino acid (Cashel et al., 1996). The amount of ppGpp produced depends upon the identity of the absent amino acid (Donini et al., 1978). Activation by ppGpp is essential to enhanced expression of the deattenuated leu operon (Wright et al., 1999), as well as a number of other amino acid operons (Morse & Morse, 1976; Perel’man & Shakulov, 1981; Smolin & Umbarger, 1975; Stephens et al., 1975; Zidwick et al., 1984). Amino acid biosynthetic genes regulated by repression are also activated by ppGpp in relA+ cells (Zidwick et al., 1984). The arginine regulon consists of 11 genes, all of which are repressed at the operator site by arginine when 1457

J. M. Reimers and others

combined with the protein encoded by argR. Positive regulation by ppGpp occurs at both the transcription and the translation level (Williams & Rogers, 1987; Zidwick et al., 1984). Starving an auxotroph for its required amino acid will derepress/deattenuate the operon (including the defective gene) and reversion rates to wild-type should increase if transcription enhances mutations. Cells with the ability to activate transcription in the presence of ppGpp (relA+) would be expected to have higher rates of transcription and mutation than cells that can not activate derepressed genes (relA mutants). In addition, inactivation of a repressor would also be expected to increase transcription and thereby affect mutation rates. Unlike the situation in regulated strains, mutations could occur during growth in repressor knockouts because transcription is not repressed by the presence of the required amino acid. Previous investigations with multiple auxotrophs of E. coli (CP78 and CP79, which are isogenic except for relA) have demonstrated a positive correlation between reversion rates in the leuB gene, ppGpp levels, and the concentration of leuB mRNA after 15 min of leucine starvation (Wright et al., 1999; Wright & Minnick, 1997). Increases in leuB and argH mRNA levels are known to be specific to the starvation conditions, i.e. leuB mRNA accumulates during leucine, but not arginine or threonine starvation, and argH mRNA accumulates during arginine, but not histidine starvation (Wright et al., 1999). These observed changes in mRNA levels in response to nutritional stress could result from a change in the rate of mRNA synthesis, degradation (stability), or both. It is commonly assumed that mRNA concentration reflects the rate of transcription. This relationship has been documented in a number of studies in which transcription increased in response to environmental conditions (Meyer & Schottel, 1991; Pease & Wolf, 1994). However, very few mRNA stability studies have been published. In fact, Bernstein et al. (2002) found published reports of RNA half-life for less than 0?5 % of the 4288 predicted ORFs in the E. coli genome. This small number of investigations could reflect the costly and labour-intensive nature of such work. Several of these mRNA stability studies do reveal that high levels of gene expression are due either solely or partially to increased stability of the mRNA (Bricker & Belasco, 1999; Georgellis et al., 1993; Zgurskaya et al., 1997). Currently, DNA microarray studies are widely used to indicate changes in gene expression when cells are exposed to a new condition or event. While microarrays measure mRNA abundance, they should not be used as an indication of gene ‘expression’ unless combined with a study to determine if the changes in quantity are due to increased stability of the message or increased rates of mRNA synthesis. In this study, mRNA concentrations and half-lives were analysed and transcription rates were calculated at several 1458

points during growth and starvation to determine whether increased stability, increased synthesis or both accounted for the increase in concentration of mRNA. These transcription rates were then compared with mutation rates to determine if increased synthesis of mRNA is correlated with increased mutation rates.

METHODS Bacterial strains and growth conditions. The bacterial strains,

CP78 and CP79, used in this study are multiple auxotrophs of E. coli K-12, requiring leucine, arginine, threonine and histidine (Wright & Minnick, 1997). They are isogenic, except that CP78 is relA+ and CP79 is relA2. The argR mutants (DargR : : cat), DR78 and DR79 (derived from CP78 and CP79 respectively) were obtained by homologous recombination using the pWM91 suicide vector (Metcalf et al., 1996) containing the entire argR gene disrupted with a cat cassette. Successful recombination resulted in chloramphenicolresistant cells. PCR analysis verified the presence of the cat cassette within the argR gene. Derepression of the argCBH genes was confirmed by comparing the activity of the argB enzyme in our argR knockout strain to that of a known argR mutant (MA1030: CGSC no. 1184). Cells were grown as previously described at 37 uC (Wright & Minnick, 1997). Growth was monitored by following the OD550 of the cultures (Perkin Elmer model 35 spectrophotometer). Total RNA was isolated during growth by removing aliquots from exponentially growing cultures. For isolation of RNA during starvation, cells were washed out of minimal medium by centrifuging for 1 min at 10 000 g and resuspended in pre-warmed 37 uC minimal medium lacking leucine or arginine. Starved cultures were incubated as before and aliquots were taken at specified time points. Mutation rate determinations. Conditions for growth and deter-

mination of mutation rates have been previously described (Wright, 1996; Wright & Minnick, 1997; Wright et al., 1999). A large culture was inoculated with cells from a 7 h old nutrient agar plate and 1?5 ml aliquots were distributed into 40 2 cm diameter test tubes, which were shaken at a 45u angle at 37 uC until the supply of the limiting amino acid was exhausted and growth ceased. Each entire culture was plated onto selective medium and incubated at 37 uC. Several identical cultures were diluted and plated onto nutrient agar plates to determine viable cell numbers. The number of selective plates without revertants was counted after 48, 65 and 72 h. Mutation rates were estimated by the ‘zero method’ of Luria & Delbru¨ck (1943) according to the expression MR = (2ln2) (ln P0/N), where P0 is the proportion of cultures with no revertants, and N is the total number of cells per culture. The incubation period chosen for each strain on selective media depended upon a number of variables. For the argH reversions a compromise (i.e. 48 h) was necessary to read the plates when the majority of true revertants were expressed while the minimum numbers of suppressors were observed. In the case of leuB reversions, suppressors were negligible and at 65 h of incubation, 90 % of the revertants expressed were true. To correct for the presence of intergenic suppressors of the argH mutation, at least 20 revertants from each strain were isolated at various times after plating. These were sequenced to find the ratio of true revertants to suppressors. The percentage of true revertants was used to correct the P0 component of each mutation rate calculation. Sequencing showed that the leuB mutation was a C-to-T transition changing TCG (serine) to TTG (leucine), and revertants were either TCG (serine) or GTG (valine) (Wright & Minnick, 1997). The argH mutation was a G-to-A transition changing TGG (tryptophan) to TGA (stop). The revertants appearing most often were TGG (tryptophan), Microbiology 150

Transcription rates and mutation in E. coli TTA (leucine), TGC (cysteine), AGA (arginine) and CGA (arginine). TGT (cysteine), TCA (serine) and GGA (glycine) were observed less frequently. RNA isolation. Total cellular RNA was isolated according to the hot phenol method of Tsui et al. (1994). Cells were added directly to lysis buffer at 100 uC, extracted twice with phenol at 65 uC, once with phenol/chloroform/isoamyl alcohol (25 : 24 : 1, by vol.) at room temperature, and once with chloroform/isoamyl alcohol (24 : 1, v/v). After two diethyl ether extractions, RNA was precipitated with 2-propanol at 220 uC overnight. This pellet was washed with 70 % ethanol, resuspended in water and precipitated with 0?5 vol. 7?5 M LiCl at 220 uC. After a 70 % ethanol wash, the RNA was treated with 5 U RQ1-RNase-free DNase (Promega) for 15 min at 37 uC. DNase was removed by phenol/chloroform/isoamyl alcohol extraction followed by chloroform extraction. The RNA was precipitated with 0?1 vol. 5 M ammonium acetate and 2?5 vols 100 % ethanol overnight at 220 uC. After centrifugation at 16 000 g for 20 min at 4 uC, the final RNA pellet was resuspended in 200 ml water. Total RNA concentration was determined by reading the A260 (Beckman DU 650 spectrophotometer). S1-nuclease protection assays. A 4- to 10-fold molar excess of biotinylated antisense RNA probe (Wright et al., 1999), 0?5–100 mg total cellular RNA and 10 mg yeast tRNA were precipitated with 0?1

vol. 5 M ammonium acetate and 3 vols 100 % ethanol at 220 uC for 30 min. After centrifugation at 16 000 g for 15 min at 4 uC, the pellets were air-dried for 30 min and resuspended in 10 ml RPAII hybridization buffer (Ambion). The samples were hybridized overnight at 50 uC for leuB or 52 uC for argH. Unhybridized RNA and probe were digested with 300 U S1-nuclease (Promega) at 42 uC for 45 min. The digest was stopped by adding 0?2 vol. stop buffer [4 M ammonium acetate, 30 mM EDTA, 170 mg tRNA ml21 and 0?7 mg GlycoBlue ml21 (Ambion)] and the protected hybrids were precipitated with 2?5 vols 100 % ethanol. After precipitation at 220 uC for 30 min and centrifugation at 16 000 g for 15 min at 4 uC, the pellets were air-dried for 30 min and suspended in 11 ml Gel Loading Buffer II (Ambion). The samples were boiled for 4 min, loaded onto pre-run 5 % (w/v) polyacrylamide gel (19 : 1, w/w, acrylamide/bisacrylamide) containing 8 M urea and 16 TBE and electrophoresed at 40 V cm21 until the bromphenol blue band ran off the gel. The gels were electroblotted onto nylon membranes; the hybrids were cross-linked to the membrane with UV light, washed and bound with streptavidin–alkaline phosphatase. Chemiluminescent reagent, CDP-Star (Ambion), was added and hybrids were detected with X-ray film as described previously (Wright et al., 1999). Scanning densitometry and comparison to known amounts of standards were used to determine specific mRNA concentrations. Half-life determinations. Decay rates of specific mRNAs, at speci-

fic times of starvation, were measured by inhibiting transcription initiation with 300 mg rifampicin ml21. Aliquots of cells were removed immediately before and at several time points after rifampicin addition. Total cellular RNA was extracted, hybridized to gene-specific probes and the amounts of specific mRNAs were determined by S1-nuclease protection assays. The half-life was calculated from the slope of a least-squares regression line of a semilogarithmic plot of percentage mRNA remaining as a function of time. The half-life was calculated using time points determined when the decay rate was exponential. Transcription rates. Rates of mRNA synthesis (KS) were calculated

from the measured values of mRNA concentration ([mRNA]) and halflife (t1/2) according to the following equation: KS=(ln2 [mRNA])/t1/2 (Zgurskaya et al., 1997). Determination of ppGpp. Cells were grown in minimal medium

as previously described (Wright, 1996) with 0?05 M MES (pH 6?5) http://mic.sgmjournals.org

and 0?05 M KH2PO4 substituted for 0?04 M sodium phosphate buffer (pH 6?5). When cells were to be starved for an amino acid, it was present at 0?2 mM in the growth medium. When cultures reached an OD550 of 0?1, they were labelled with 100 mCi (3?7 MBq) H332PO4 ml21 (ICN; 3?76106 Bq ml21) and then grown at 37 uC to an OD550 of 0?3. Growth was monitored in an unlabelled companion culture. The cells were then starved for either arginine or leucine by centrifuging at 16 000 g for 1 min, and resuspending the cell pellet in radioactive minimal medium lacking arginine or leucine. Immediately before and at several time points during starvation, 50 ml aliquots of cells were removed and added to 17 ml 23?6 M formic acid on ice, quickly mixed and frozen in a dry ice/ ethanol bath. When samples were collected at all time points, the samples were freeze–thawed twice and centrifuged for 1 min. Aliquots (2 and 5 ml) were spotted onto washed PEI cellulose plates. Forty nanomoles of GTP and ppGpp (Trilink Biotechnologies) treated in the same manner were spotted as standards. Ascending chromatography in 1?5 M KH2PO4 (pH 3?4) was run at 5 uC until the solvent front was 16 cm from the origin. The radioactive spots were located by autoradiography. The radioactive areas corresponding to ppGpp were excised from the plate, wetted with 0?25 ml water and their radioactivity measured in 5 ml Ecolite (ICN). The concentration of ppGpp [nmol (OD550 unit)21] was based on the specific radioactivity of the phosphate in the medium (Wright & Minnick, 1997).

RESULTS AND DISCUSSION In the leu operon and arg regulon, reversions of mutants would not be expected to occur during growth in the presence of their end product, when transcription rates are low. During starvation, when transcription is activated by deattenuation or derepression, reversion rates should increase. In contrast, in the arginine repressor knockout strains (unregulated), reversions could occur during growth in the presence of arginine, because transcription rates are high due to the absence of the repressor protein. The uniform size of revertant colonies formed on selective plates indicated that mutations occurred more or less synchronously. Although this suggests that the reversions occurred over a brief period of time, the exact time is unknown. Therefore, transcription rates were calculated for periods encompassing the most likely times reversions could occur, namely: (1) growth, (2) peak mRNA synthesis and (3) starvation steady state. In regulated strains, transcription during growth would not be expected to contribute significantly to mRNA synthesis. However, in unregulated strains, transcription during growth would be expected to make a significant contribution to the overall transcription rate. In order to compare transcription rates, the same time periods of growth and starvation were analysed for all strains. During experiments to determine mRNA concentrations and half-lives, cells grew exponentially for approximately 4 h before enough cells were present to harvest or to initiate starvation experiments. However, when calculating transcription rates, a more conservative estimate of 2 h for the duration of exponential growth was chosen. The effects of starvation were very dynamic in the first few minutes, but after 30 min mRNA levels appeared to approximate a steady state. Therefore, during starvation, mRNA 1459

J. M. Reimers and others

The concentration of leuB mRNA in CP78 increased 60fold above the level during growth by 10 min of starvation, then declined to a level 24-fold of that during growth and remained at that level for at least 20 min. Preliminary experiments (not shown) indicated that the level of mRNA reached by 30 min of starvation was maintained for approximately 8 h. In CP79 the concentration of leuB mRNA was increased 45-fold by 4 min of starvation, after which there was a rapid decline, and very little mRNA remained after 15 min.

Fig. 1. Concentration of leuB mRNA (a) and ppGpp (b) in CP78 ($) and CP79 (#) strains during growth and starvation for leucine. Leucine was removed from cell cultures at 0 min (see Methods). The mRNA concentration data are the means of two to four independent experiments and the error bars indicate SD. The ppGpp concentration data represent the means of duplicate samples which differed by no more than 20 %.

concentrations and half-lives were measured at peak mRNA levels and at the 30 min starvation steady state. Assuming that reversions occurred within this 2?5 h period of growth plus starvation, transcription rates and amounts of mRNA produced during each phase were determined and a mean rate of transcription for each gene was calculated. Levels of leuB mRNA and ppGpp in CP78 and CP79 Concentrations of leuB mRNA were measured every few minutes during growth and starvation for leucine (Fig. 1a).

In CP78, the increase in leuB mRNA levels was slightly preceded by and correlated with increased ppGpp levels, as previously reported, while CP79 (relA2) produced minimal amounts of ppGpp (Fig. 1b; Donini et al., 1978). Deattenuation was probably responsible for operon expression during the first few minutes of leucine starvation, since leuB mRNA accumulated immediately in both CP78 (+ppGpp) and CP79 (2ppGpp). However, as starvation continued, only CP78 synthesized leuB mRNA, and by 20 min of starvation, 96 % of leuB mRNA synthesis was ppGpp-dependent. CP78, with its higher levels of ppGpp, was able to maintain higher levels of mRNA for a much longer time than CP79 (Fig. 1a, Table 1). Half-lives and synthesis rates of leuB mRNA in CP78 and CP79 S1-nuclease protection assays were used to determine mRNA concentrations and decay rates in the presence of rifampicin (1) during exponential growth, (2) at peak mRNA concentration during leucine starvation, and (3) after 30 min of leucine starvation, during the new starvation steady-state. Transcripts of leuB became two to three times more stable during starvation (Fig. 2a, b, Table 1) and CP79 leuB mRNA showed an increase in stability over CP78. Transcription rates were calculated (see Methods) using the concentration and half-life data obtained at each of the above time points. A low level of mRNA synthesis was maintained during growth in both CP78 and CP79 (Table 1). Within 10 min of leucine starvation, there was a 60-fold increase in synthesis rate in CP78. After this peak level was reached, the synthesis rate decreased, but CP78

Table 1. Concentrations, half-lives and synthesis rates of leuB mRNA during growth and during leucine starvation The values given are the means of three to six independent determinations±SD; see Methods for details. Strain

Relevant genotype

CP78

leuB6 relA+

CP79

1460

leuB6 relA2

Time of determination Growth Starved for Starved for Growth Starved for Starved for

10 min 30 min 4 min 30 min

leuB mRNA concn [pg leuB mRNA (mg total RNA)”1]

Half-life (min)

Synthesis rate [pg leuB mRNA (mg total RNA)”1 min”1]

8±2 485±29 193±20 7±3 319±70 10±4

1?1±0?4 1?1±0?3 1?8±0?4 1?0±0?2 1?9±0?2 2?8±0?5

5 306 74 5 116 2

Microbiology 150

Transcription rates and mutation in E. coli

Fig. 2. Decay of mRNA during exponential growth (m), at peak mRNA concentration ($) and at 30 min of starvation (&), for leuB mRNA in CP78 (a) and CP79 (b), and for argH mRNA in CP78 (c), CP79 (d), DR78 (e) and DR79 (f). The data represent the means of two or three samples which differed by no more than 20 %. The half-life values determined from these data are shown in Tables 1 and 3.

continued to transcribe at a rate 15-fold higher than that during growth. In CP79, the rate of synthesis reached a maximum at 4 min of starvation, but rapidly declined to a level equivalent to that during growth. Even at its maximal rate, CP79 synthesized mRNA at one-third the rate of CP78 (Table 1). Correlation of leuB transcription with reversion rates The mean leuB mRNA transcription rate (Table 2) during the period that the reversions were expected to occur was calculated by determining the amount of mRNA produced during growth (2 h) and starvation (30 min) and dividing by the total time (150 min). The amount of mRNA produced during each time period was calculated by determining the time each rate was in effect from the profiles of http://mic.sgmjournals.org

the graphs in Fig. 1(a) and multiplying by the transcription rate during each period. The mean rate of leuB mRNA synthesis was fourfold higher in CP78 than in CP79 (Table 2). Reversion rates for amino acid auxotrophs were determined previously (Wright, 1996) by growing the cells to stationary phase in a limiting amount of the required amino acid and then plating to minimal medium lacking that amino acid. Only those cells that revert to prototrophy survive. When cells were starved for leucine, the reversion rate of leuB6 in CP78 was sevenfold higher than that in CP79 (Table 2; Wright, 1996). By comparing strains isogenic except for relA, we have shown that the strain able to activate leuB transcription with ppGpp had a higher mutation rate. Table 2 also shows the correlation between leuB reversion rates and transcription rates during the 2?5 h 1461

J. M. Reimers and others

Table 2. leuB mRNA synthesis rates and mutation rates in CP78 and CP79 Synthesis rates from Table 1, duration of rate from Fig. 1(a) and mutation rates from Wright (1996). Strain

leuB mRNA synthesis rate [pg leuB mRNA (mg total RNA)”1 (min)”1]

Duration of rate (min)

Amount of leuB mRNA produced [pg leuB mRNA (mg total RNA)”1]

CP78

5 306 74

122* 8 20 150 122* 4 24 150

610 2448 1480 4538 610 464 48 1122

Total CP79

Total

5 116 2

Mean transcription rate [pg leuB mRNA (mg total RNA)”1 min”1]

1096Mutation rate (reversions per cell per generation)

30?3

1?5±0?5

7?5

0?22±0?08

*Including the 2 min transition time between growth and starvation.

period of analysis. There was a sevenfold increase in mutation rate in CP78 (relA+) and a fourfold increase in transcription rate when compared to CP79 (relA2). Levels of argH mRNA and ppGpp in CP78 and CP79 The concentration of argH mRNA in CP78 and CP79 was analysed during growth and during 30 min of arginine starvation (Fig. 3a). In CP78, argH mRNA concentration

increased 200-fold after 12 min of arginine starvation and then gradually declined to a level 78-fold higher than during growth (Table 3). In CP79, argH mRNA concentration increased 80-fold after 12 min of arginine starvation and continued to rise, until by 30 min it was 132-fold higher than during growth (Table 3). This result agrees with previous investigations, in which the amount of argCBH mRNA was higher in a relA mutant after 30 min of arginine downshift (Williams & Rogers, 1987).

Fig. 3. (a, b) Concentration of argH mRNA (a) and ppGpp (b) in CP78 (m) and CP79 (n) during growth and starvation for arginine. (c, d) Concentration of argH mRNA (c) and ppGpp (d) in DR78 (&) and DR79 (%) during growth and starvation for arginine. Arginine was removed from cell cultures at 0 min (see Methods). The mRNA concentration data are the means of two to four independent experiments and the error bars indicate SD. The ppGpp concentration data represent the means of duplicate samples which differed by no more than 20 %. 1462

Microbiology 150

Transcription rates and mutation in E. coli

Table 3. Concentrations, half-lives and synthesis rates of argH mRNA during growth and during arginine starvation in argR+ and argR knockout strains The values given are the means of three to six independent determinations±SD; see Methods for details. Strain

Relevant genotype

CP78

argH46 relA+

CP79

DR78

DR79

argH46 relA2

argH46 relA+ DargR : : cat argH46 relA2 DargR : : cat

Time of determination Growth Starved for Starved for Growth Starved for Starved for Growth Starved for Starved for Growth Starved for Starved for

argH mRNA concn [pg argH mRNA (mg total RNA)”1]

12 min 30 min 12 min 30 min 12 min 30 min 12 min 30 min

0?4±2 81±13 31±13 0?5±1 40±14 66±13 76±13 34±8 23±7 61±12 43±10 51±12

Half-life (min)

Synthesis rate [pg argH mRNA (mg total RNA)”1 min”1]

ND

(1)D 56 24 (1)D 14 38 105 34 23 70 30 17

1?0±0?2 0?9±0?4 ND

2?0±0?3 1?2±0?4 0?5±0?2 0?7±0?3 0?7±0?1 0?6±0?2 1?0±0?1 2?1±0?4

ND, Not determined due to low concentration of argH mRNA during growth. DA synthesis rate of 1 pg argH mRNA (mg total RNA)21 min21 was estimated; see text.

The increase in argH mRNA at 12 min of arginine starvation (Fig. 3a) correlated with the increase in ppGpp in the relA+ strain at 12 min of starvation (Fig. 3b). This is consistent with the results of Donini et al. (1978) showing that ppGpp levels increased in CP78 at about 8 min of arginine starvation, but only minimally in CP79. Since significant levels of ppGpp were only present in CP78, derepression and/or a decrease in mRNA half-life was apparently responsible for the increased concentration of argH mRNA in CP79. Half-lives and synthesis rates of argH mRNA in CP78 and CP79 Half-lives and concentrations of argH mRNA were measured (1) during growth, (2) at peak mRNA synthesis during arginine starvation and (3) after 30 min of arginine starvation in CP78 (Fig. 2c) and in CP79 (Fig. 2d). The concentration of argH mRNA during exponential growth was too low to determine its decay rate. However, during arginine starvation, when half-lives could be measured, there was no change in half-life in CP78 (Fig. 2c, Table 3), while argH mRNA became slightly less stable in CP79 (Fig. 2d, Table 3). Half-lives and concentrations of argH mRNA determined above were used to calculate rates of transcription during each time period (Table 3). Very low argH mRNA concentrations did not allow for a synthesis rate to be determined during exponential growth. However, since this concentration was extremely low, the rate of synthesis could not be significant even if the half-life were very short. For example, at an mRNA concentration of 0?5 pg mRNA (mg total RNA)21, even with a very short half-life of 0?3 min (one-third of that measured during arginine http://mic.sgmjournals.org

starvation) the synthesis rate would be only 1 pg mRNA (mg total RNA)21 min21. This rate was chosen as the highest possible estimate and used below to determine the mean transcription rate. At 12 min of arginine starvation, the synthesis rate of argH mRNA in CP78 was fourfold higher than in CP79 (Table 3). However by 30 min of starvation, CP79 was transcribing at a slightly higher rate than CP78. Correlation of argH transcription with reversion rates in CP78 and CP79 The mean synthesis rates of argH mRNA in CP78 and CP79 (Table 4) were calculated as previously described for leuB mRNA. Since an actual rate during growth could not be determined, an estimated synthesis rate of 1 pg mRNA (mg total RNA)21 min21 was used to calculate mean transcription rate. The mean transcription rate of argH mRNA was 1?4-fold higher in CP78 than in CP79. When CP78 was starved for arginine, the argH46 mutation reverted at a rate twofold higher than in CP79 (Table 4). Therefore, the relA+ strain, compared to the relA2 strain, had a higher reversion rate, indicating that the transcriptional activator ppGpp had a positive influence on the mutation of this gene. When comparing CP78 and CP79 (Table 4), there was a qualitative correlation between the twofold increase in mutation rate and the 1?4-fold increase in transcription rate of the argH gene in the relA+ strain. Reversion rates for the argH gene in CP78 and CP79 (Table 4) were determined by counting negative plates at 48 h, instead of 72 h as previously reported (Wright, 1996). Sequencing revertants that appeared early (48 h) as well as those that appeared later (72 h) showed that the 1463

J. M. Reimers and others

Table 4. argH mRNA synthesis rates and mutation rates in CP78, CP79, DR78 and DR79 Synthesis rates from Table 3; duration of rate from Figs 3(a) and (c). Mutation rate values are the means of three to eight independent determinations±SD. Strain

argH mRNA synthesis rate [pg argH mRNA (mg total RNA)”1 (min)”1]

Duration of rate (min)

Amount of argH mRNA produced [pg argH mRNA (mg total RNA)”1]

CP78

(1)* 56 24

124D 8 18 150 124D 16 10 150 124D 10 16 150 124D 16 10 150

124 448 432 1 004 124 224 380 728 13 020 340 368 13 728 8 680 480 170 9 330

Total CP79

Total DR78

Total DR79

Total

(1)* 14 38 105 34 23 70 30 17

Mean transcription rate [pg argH mRNA (mg total RNA)”1 (min)”1]

1096Mutation rate (reversions per cell per generation)

6?7

0?26±0?2

4?9

0?12±0?08

91?5

6?9±1?5

62?2

3?2±1?2

*A synthesis rate of 1 pg argH mRNA(mg total RNA)21 min21 was estimated; see text. DIncluding the 4 min transition time between growth and starvation.

colonies appearing after 48 h were largely suppressors. Since the argH46 mutation creates a stop codon, it is likely that these later colonies were intergenic suppressors. Twenty early (48 h) revertants of each strain were sequenced and the percentages of true revertants were determined. In CP78, 89 % of the scored revertants were true revertants, while only 30 % of the revertants scored in CP79 were true. Sequencing was necessary since colony size did not indicate whether the colony came from a true revertant or a suppressor. These percentages were used to calculate the reversion rates reported in Table 4. Levels of argH mRNA and ppGpp in the argR knockout strains DR78 and DR79 The argR gene in CP78 and CP79 was inactivated by the insertion of the cat gene, producing the strains DR78 (relA+ argH46 DargR : : cat) and DR79 (relA2 argH46 DargR : : cat). Since the normal transcriptional control was removed in these knockout strains, unregulated (increased) transcription would be expected to occur during growth in the presence of arginine. Cunin et al. (1969) reported that in the absence of the repressor, the enzymes encoded by the argCBH transcriptional unit increased 60-fold in activity. Unlike the regulated (repressed) strains, high argH mRNA levels were found during exponential growth in both DR78 and DR79. In fact, the argR knockout strains (Fig. 3c, Table 3) had 100- to 200-fold more argH mRNA than CP78 and CP79 (Fig. 3a, Table 3). The argH mRNA 1464

levels declined as arginine starvation became more severe, and after 30 min of starvation, the level of argH mRNA was higher in DR79 than in DR78. Interestingly, peak concentrations of mRNA during growth of the argR knockout strains (Fig. 3c) are similar to peak values during starvation in the regulated strains (Fig. 3a). This suggests that removing the ArgR protein fully derepressed the argH gene. In the argR knockout strains, DR78 and DR79, ppGpp concentrations were analysed during arginine starvation. In DR78 (relA+), ppGpp accumulated (Fig. 3d) while DR79 (relA2) produced minimal amounts of ppGpp when starved for arginine. Comparing Figs 3(c) and 3(d), there did not appear to be a correlation between ppGpp levels and mRNA levels; but as discussed below, the large increase in transcription due to the inactivation of the repressor perhaps masked a relatively small activation by ppGpp. Half-lives and synthesis rates of argH mRNA in DR78 and DR79 In DR78, argH mRNA half-lives remained short during arginine starvation (Fig. 2e, Table 3), but in DR79 half-lives increased threefold (Fig. 2f, Table 3). During exponential growth in both DR78 and DR79, both concentrations and synthesis rates were very high (Table 3). As starvation continued the argH mRNA concentration and its synthesis rate decreased. At 30 min of starvation, even though the mRNA concentration was twofold higher in DR79, mRNA Microbiology 150

Transcription rates and mutation in E. coli

synthesis was lower than that in DR78 because the mRNA half-life in DR79 had increased threefold. Thus, in the case of DR78 and DR79, increased stability did contribute to the observed increase in concentration and therefore, mRNA concentration was a poor indicator of the rate of transcription. Correlation of argH transcription with reversion rates in DR78 and DR79 The mean synthesis rate of argH mRNA in DR78 and DR79 was calculated as previously described and is shown in Table 4. The mean transcription rate in DR78 was 1?5-fold higher than in DR79 (Table 4); therefore ppGpp in DR78 had a similar positive effect on transcription as in CP78. During growth, the ArgR protein, in combination with arginine, represses transcription of argH. If increased transcription increases mutation rates, then the reversion rate of argH should be higher in the (unregulated) argR mutants than in the argR+ strains. The argR knockout strains (DR78 and DR79) had the highest observed mutation rates, 6?961029 reversions per cell per generation in DR78 and 3?261029 in DR79 (Table 4). In DR78, 100 % of the revertants were true and in DR79, 72 % of the revertants were true. Perhaps the higher percentage of true revertants in the relA+ strains, as well as in the argR mutants, reflected a higher reversion rate due to the higher rate of transcription. Table 4 shows the correlation between rates of argH transcription and reversion rates in the argR mutants. The transcription rate of the relA+ strain increased 1?5-fold over that of the relA2 strain, while the mutation rate increased approximately twofold. Another set of correlations can be observed when comparing the regulated (CP78, CP79) and unregulated (DR78, DR79) strains. Comparing the argR+ and argR knockout strains, there was a 26-fold increase in mutation rates in the absence of the repressor for both CP78 versus DR78 and CP79 versus DR79. There was also a 13-fold increase in mean transcription rates when comparing CP78 with DR78 and CP79 with DR79. Therefore, both transcription and mutation rates increased by an order of magnitude when the regulated and unregulated strains were compared. These knockout strains, lacking the repressor protein, showed the greatest effect on argH mRNA synthesis and the highest increase in mutation rate yet observed. They illustrate most clearly that increasing transcription increases mutation rates. It was surprising that in the two genes and four strains used here, there is a single linear relationship between increased transcription and reversion rates (Fig. 4). While it is to be expected that such a correlation may exist for a single gene when its level of transcription is varied, it was not expected when comparing different genes due to the many other variables impacting on both transcription and mutation rates. For example, a gene’s transcription rate is http://mic.sgmjournals.org

Fig. 4. Correlation of reversion rates (reversions per cell per generation) with transcription rates [pg mRNA (mg total mRNA)”1 (min)”1] for the leuB (n) and argH ($) mutants; R=0?98, P=0?02. Reversion rate data are the mean of three to eight independent experiments. Transcription rate data are from Tables 2 and 4.

affected by DNA-binding proteins and the level of supercoiling in its immediate environment. However, the correlation seen in Fig. 4 suggests that variables such as these may be similar for the leuB and argH genes. This investigation, using mRNA half-life and concentration data to calculate rates of transcription, established that increased transcription rather than increased mRNA stability was primarily responsible for the increased mRNA concentrations observed. The removal of the ArgR repressor served to increase both transcription and mutation rates by more than 10-fold while ppGpp had a smaller, but consistent effect. The results of these analyses support the hypothesis that increased mutation rates in the leuB and argH genes are primarily due to increased rates of mRNA synthesis.

ACKNOWLEDGEMENTS We thank Greg St George for helpful comments on the manuscript. This work was supported by the Eppley Foundation, the NSF EPSCoR program, the Stella Duncan Research Institute and NIH grants R15CA88893 and R55CA99242.

REFERENCES Bernstein, J. A., Khodursky, A. B., Lin, P.-H., Lin-Chao, S. & Cohen, S. N. (2002). Global analysis of mRNA decay and abundance in

Escherichia coli at single-gene resolution using two-color fluorescent DNA microarrays. Proc Natl Acad Sci U S A 99, 9697–9702. Bricker, A. L. & Belasco, J. G. (1999). Importance of a 59 stem-loop

for longevity of papA mRNA in Escherichia coli. J Bacteriol 181, 3587–3590. Cashel, M., Gentry, D. R., Hernandez, V. D. & Vinella, D. (1996). The

stringent response. In Escherichia coli and Salmonella Typhimurium: Cellular and Molecular Biology, 2nd edn, vol. 1, pp. 1458–1496. 1465

J. M. Reimers and others Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology. Cunin, R., Elseviers, D., Sand, G., Freundlich, G. & Glansdorff, N. (1969). On the functional organization of the argECBH cluster of

Stephens, J. C., Artz, S. W. & Ames, B. (1975). Guanosine 59diphosphate 39-diphosphate (ppGpp): positive effector for histidine operon transcription and general signal for amino acid deficiency. Proc Natl Acad Sci U S A 72, 4389–4393.

genes in Escherichia coli K-12. Mol Gen Genet 106, 32–47.

Tsui, H. C., Pease, A. J., Koehler, T. M. & Winkler, M. E. (1994).

Donini, P., Santonastaso, V., Roche, J. & Cozzone, A. J. (1978).

Detection and quantitation of RNA transcribed from bacterial chromosomes and plasmids. Methods Mol Genet 3, 179–204.

The relationship between guanosine tetraphosphate, polysomes and RNA synthesis in amino acid starved Escherichia coli. Mol Biol Rep 4, 15–29. Georgellis, D., Barlow, T., Arvidson, S. & von Gabain, A. (1993).

Williams, M. G. & Rogers, P. (1987). Expression of arg genes in Escherichia coli during arginine limitation is dependent upon control of translation. J Bacteriol 169, 1644–1650.

Retarded RNA turnover in Escherichia coli: a means of maintaining gene expression during anaerobiosis. Mol Microbiol 9, 375–381.

Wright, B. E. (1996). The effect of the stringent response on

Luria, S. E. & Delbru¨ck, M. (1943). Mutations of bacteria from virus

Wright, B. E. (2000). A biochemical mechanism for nonrandom mutations and evolution. J Bacteriol 182, 2993–3001.

sensitivity to virus resistance. Genetics 28, 491–504. Metcalf, W. W., Jiang, W., Daniels, L. L., Kim, S.-K., Haldimann, A. & Wanner, B. L. (1996). Conditionally replicative and conjugative

plasmids carrying lacZ alpha for cloning, mutagenesis and allele replacement in bacteria. Plasmid 35, 1–14. Meyer, B. J. & Schottel, J. L. (1991). A novel transcriptional response

by the cat gene during slow growth of Escherichia coli. J Bacteriol 173, 3523–3530. Morse, D. E. & Morse, A. N. C. (1976). Dual control of the

tryptophanyl-tRNA synthetase and the repressor. J Mol Biol 103, 209–226.

mutations in Escherichia coli K-12. Mol Microbiol 19, 213–219.

Wright, B. E. & Minnick, M. F. (1997). Reversion rates in a leuB

auxotroph of Escherichia coli K-12 correlate with ppGpp levels during exponential growth. Microbiology 143, 847–854. Wright, B. E., Longacre, A. & Reimers, J. M. (1999). Hypermutation

in derepressed operons of Escherichia coli K-12. Proc Natl Acad Sci U S A 96, 5089–5094. Wright, B. E., Reimers, J. M., Schmidt, K. H. & Reschke, D. K. (2002). Hypermutable bases in the p53 cancer gene are at vulnerable

positions in DNA secondary structures. Cancer Res 62, 5641–5644.

Pease, A. J. & Wolf, R. E., Jr (1994). Determination of the growth

Wright, B. E., Reschke, D. K., Schmidt, K. H., Reimers, J. M. & Knight, W. (2003). Predicting mutation frequencies in stem-loop

rate-regulated steps in expression of the Esherichia coli K-12 gnd gene. J Bacteriol 176, 115–122.

structures of derepressed genes: implications for evolution. Mol Microbiol 48, 429–441.

Perel’man, B. V. & Shakulov, R. S. (1981). Dependence of the

Zgurskaya, H. I., Keyhan, M. & Matin, A. (1997). The ss level in

expression of the threonine operon of E. coli on the allelic state of the relA gene and the guanosine tetraphosphate content. Biokhimiya 46, 1267–1276.

starving Escherichia coli cells increases solely as a result of its increased stability, despite decreased synthesis. Mol Microbiol 24, 643–651.

Smolin, D. E. & Umbarger, H. E. (1975). Specificity of the

Zidwick, M. J., Korshus, J. & Rogers, P. (1984). Positive control of

stimulation of in vitro ribonucleic acid synthesis by guanosine 59-diphosphate 39-diphosphate. Mol Gen Genet 141, 277–284.

expression of the argECBH gene cluster in vitro by guanosine 59-diphosphate 39-diphosphate. J Bacteriol 159, 647–651.

1466

Microbiology 150