Skip to main page content

• HOME
• Current Issue
• Archive
• Alerts
• About ASM
• Contact us
• Tech Support
• Journals.ASM.org

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Buckstein, M. H. Articles by Rubin, H. Search for Related Content PubMed PubMed Citation Articles by Buckstein, M. H. Articles by Rubin, H.

 Previous Article  |  Next Article 

Journal of Bacteriology, January 2008, p. 718-726, Vol. 190, No. 2
0021-9193/08/$08.00+0     doi:10.1128/JB.01020-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Characterization of Nucleotide Pools as a Function of Physiological State in Escherichia coli

Michael H. Buckstein,¹ Jian He,² and Harvey Rubin^1*

Division of Infectious Diseases,¹ Department of Chemistry, University of Pennsylvania, Philadelphia, Pennsylvania 19104²

Received 27 June 2007/ Accepted 17 October 2007

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Using a modified method that involves minimal manipulation of cells, we report new information about nucleotide pool sizes and changes throughout the Escherichia coli growth curve. Nucleotide pool sizes are critically dependent on sample manipulation and extraction methods. Centrifugation and even short (2 min) lapses in sample preparation can dramatically affect results. The measured ATP concentration at three different growth rates is at least 3 mM, well above the 0.8 mM needed to saturate the rRNA promoter P1 in vitro. Many of the pools, including ATP, GTP, and UTP, begin to decrease while the cells are still in mid-log growth. After an almost universal drop in nucleotide concentration as the cells transition from logarithmic to stationary phase, there is a "rebound" of certain nucleotides, most notably ATP, after the cells enter stationary phase, followed by a progressive decrease. UTP, in contrast, increases as the cells transition into stationary phase. The higher UTP values might be related to elevated UDP-glucose/galactose, which was found to be at higher concentrations than expected in stationary phase. dTTP is the most abundant deoxynucleoside triphosphate (dNTP) in the cell despite the fact that its precursors, UDP and UTP, are not. All dNTPs decrease through the growth curve but do not have the abrupt drop, as seen with other nucleotides when the cells transition into stationary phase.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
It is increasingly recognized that the relative concentrations of nucleotides play important roles in prokaryotic cell regulation. The classic example of this is the stringent response in which a hyperphosphorylated guanosine, ppGpp, is produced in response to stalled ribosomes in amino acid-starved cells (8, 9, 11). ppGpp is produced even under nutrient-rich growth conditions, however, and its presence probably ensures transcriptional balance for different promoters under a variety of physiological states (13, 24, 26). It is well accepted now that ppGpp is the primary, although not the exclusive, determinant of growth rate dependence: the observation that the amount of rRNA produced in Escherichia coli is proportional to the growth rate.

In addition to alarmones such as ppGpp, the relative concentrations of even standard nucleotides such as ATP and GTP affect bacterial physiology. One theory suggests the concentration of initiating nucleotide (iNTP) for rRNA promoters changes under different growth conditions, which in turn affects the transcriptional rate from these promoters (15). Although it appears under different growth rates the concentration of ATP, the iNTP for rRNA promoter P1, does not change, it has been observed that the ATP concentration does decrease as cells enter stationary phase (21, 24). Similarly, Fis, a nucleoid-associated protein that affects transcription, is controlled at the promoter level by concentrations of its iNTP, CTP (32). The ratio of the nucleotide triphosphates (NTPs) to diphosphates also has regulatory consequences. In Bacillus subtilis, the ratio of GTP to GDP plays a critical role in the suppression/induction of sporulation through the transcriptional repressor, CodY (20, 23, 29). Also, SpoIIAB sequesters ^F, a transcription factor associated with early gene expression in sporulation, in the context of high ATP and low ADP and correspondingly releases it when the ratio is reversed (2).

Nucleotide pools in bacteria have been studied using several approaches, and much of the debate in the literature about pool fluctuations and their biological roles stems from the different methodologies used. The most traditional approach involves growing bacteria in a medium containing radioactive phosphate, extracting the nucleotides, followed by analysis with known standards on two-dimensional thin-layer chromatography (2D-TLC) (4). The pool concentrations obtained using this method are generally accepted as the gold standard. However, separations on 2D-TLC plates are not absolutely quantitative and separation of certain nucleotides such as the deoxy-NTPs (dNTPs) from their ribose counterparts is difficult.

Methods using high-pressure liquid chromatography (HPLC) have also been developed to study nucleotide pools (19, 25). However, in order to generate enough material for analysis, large cultures must be used, and some cellular manipulation such as vacuum filtration, formaldehyde fixation, and/or pelleting is needed to concentrate the samples into an appropriate volume prior to extraction. In addition, the method of extraction, acidic or basic, can produce very different results. Therefore, it is recognized that a new comprehensive HPLC-based analysis is needed to accurately characterize nucleotide pools (13).

We addressed these problems by modifying an HPLC-based acid extraction method for analyzing intracellular nucleotide pools in bacteria that is sufficiently quantitative to settle outstanding questions in the literature. In doing so, this work confirms the extraordinary lability of the nucleotide pools and the need for careful attention to preparative steps when analyzing these molecules in order to capture true reflections of pool sizes. Using the new method, we also present a nucleotide profile of E. coli as it transitions from exponential growth to early stationary phase, showing unexpected behavior of common nucleotides and high concentrations of molecules such as UDP(G) in the early stationary phase.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
All experiments were done with E. coli MG1655 (Fig. 1). Cells were cultured in modified forms of morpholinepropanesulfonic acid (MOPS) medium with 1 mM phosphate and 10 µg of thymine/ml and either 0.4% glycerol, 0.4% glucose, or 0.4% glucose and 80 mg of all 20 amino acids/ml. Starter cultures were grown overnight in 0.4% glucose MOPS and diluted 200-fold in the desired medium. Cultures were grown at 37°C with shaking until the desired optical density at 600 nm (OD₆₀₀) was obtained. Cells were collected at each time point in order to obtain approximately 1 x 10¹⁰ to 5 x 10¹⁰ cells per sample; this translates into roughly 50 ml of culture at an OD₆₀₀ of 1.0. Although fewer cells can be used, this amount allows for large peaks on HPLC and extra sample for multiple analyses. The appropriate cell volume was then poured into formic acid to a final concentration of 1 M. The sample was immediately frozen via liquid nitrogen or dry ice-ethanol. Large-volume samples required at least 2 min to freeze but were cooled within 30 s. We saw no substantial changes in nucleotide concentrations when the samples were not rapidly frozen but continued to freeze the samples for ease of being able to delay nucleotide purification until the following day and to preserve consistency in sample treatment. Although not examined further, samples could remain frozen for at least 1 week without changes in the results.

View larger version (20K): [in this window] [in a new window]

FIG. 1. Flow chart of method.

For the extraction, the cells were rapidly thawed in a 37°C water bath and kept on ice with frequent vortexing for 30 min. The cell debris was pelleted at 7,000 x g for 7 min, and the supernatant was collected and filtered to remove residual cellular debris prior to loading on the collecting column. Nucleotides were collected from the medium using Q-Sepharose Fast Flow (Amersham) in a 15-mm-diameter column with a bed volume of 3 ml. Prior to loading onto the column, the sample was diluted 20-fold in water to bring the salt concentration sufficiently low for binding. The extraction mixture was gravity dripped over the column to collect the nucleotides at a flow rate of approximately 3 to 5 ml/min. Up to 2 liters can be passed over the column before nucleotides are lost based on flow. The nucleotides were eluted from the column by using 1 M ammonium formate.

In order to remove the high salt concentration used to elute from the Sepharose column, the sample was dialyzed in serial batches against 1 M sucrose for approximately 40 to 45 h by using 100-molecular-weight-cutoff dialysis tubing (Spectra/Por). Although a minimum time for dialysis was not determined, 40 h conveniently allowed the sample to be dialyzed for 2 nights. The sample was collected from the dialysis tubing, frozen, and lyophilized. Before injection into the HPLC apparatus, the sample was dissolved in 500 µl of distilled water.

HPLC was performed by using either reversed-phase or anion-exchange chromatography using an Agilent 1100 high-pressure liquid chromatograph with UV detection at 254 nm. For reversed-phase chromatography, nucleotides were separated on a Waters Symmetry C₁₈ 3.5 µM (150 by 4.6 mm) column equipped with a NovaPak C₁₈ Sentry (Waters) guard column, adapting a prior method used for the separation of nucleotides (16). At a flow rate of 0.8 ml/min, a linear gradient of 70:30 buffer A to buffer B was run to 40:60 at 30 min. The gradient was then changed from 40:60:0 to 0:70:30 for buffer A-buffer B-buffer C. Buffer A consisted of 5 mM t-butyl ammonium phosphate (PicA Reagent; Waters), 10 KH₂PO₄, and 0.25% methanol adjusted to pH 6.9. Buffer B consisted of 5 mM t-butyl ammonium phosphate, 50 mM KH₂PO₄, and 30% methanol (pH 7.0). Buffer C was acetonitrile. For anion-exchange chromatography, a Waters Spherisorb S5 SAX (4.6 by 250 mm) column was used equipped with a Nova-Pak Silica Guard column (Waters). At a flow rate of 1.0 ml/min, a linear gradient of 100:0 to 0:100 (A:B) was run over 30 min. From 30 to 45 min, the flow rate was linearly increased to 1.5 ml/min. Buffer A consisted of 0.05 M ammonium phosphate (pH 3.4), and buffer B was 0.5 M ammonium phosphate (pH 3.4). Nucleotide standards were obtained from Sigma. ppGpp and pppGpp were generated and purified in our lab as described previously (3).

Mass measurement was performed by using a Voyager DE-RP matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometer. A total of 1 µl of matrix solution (a saturated solution of -cyano-4-hydroxycinnamic acid in 50:50 H₂O-acetonitrile) was pipetted to the stainless steel sample plate with 1 µl of unknown sample collected from the HPLC. The sample was mixed by inversion and allowed to air dry. Adjustments in the injection volume were made to account for dilution. The measurement was taken using negative ion mode with 25 kV accelerating voltage, a 95% grid voltage, and a 100-ns delay time.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Method validation. In order to analyze the complete nucleotide profile via HPLC, both anion-exchange and reversed-phase chromatography were carried out, adapting different methodologies to optimize gradients. Nucleotide standards were run on both columns. The ppGpp standard was synthesized in our lab using purified Rel_Mtb with GDP as a substrate, and its run time on each column was determined independently (3). The SAX Partisil column had a shorter run time and complete separation of all of the NTPs and ppGpp but did not resolve the dNTPs well (Fig. 2A). In contrast, the C₁₈ column did not resolve UTP from GTP and merged ppGpp with ATP (not shown in standards) but did adequately separate the dNTPs from their ribose counterparts (Fig. 2B). Recovery from each column was confirmed by running samples in parallel. There was roughly a 5 to 10% discrepancy between the columns for most nucleotides (Table 1). UTP and GTP ran together on the C₁₈ column, giving a 38% discrepancy between columns, but, when taken together, the difference between the columns was ca. 10% as well.

View larger version (21K): [in this window] [in a new window]

FIG. 2. Nucleotide standards (NDPs, dNTPs, NTPs, and pppGpp). Nucleotide standards were purchased from Sigma, with the exception of pppGpp, which was synthesized and purified in our laboratory using RelMtb (3). (A) For anion- exchange chromatography, nucleotide standards were run on a Partisil S5 SAX column at a flow rate of 1.0 ml/min. A linear gradient of 100:0 to 0:100 (A:B) was run over 30 min. From 30 to 45 min, the flow rate was linearly increased to 1.5 ml/min. Buffer A consisted of 0.05 M ammonium phosphate (pH 3.4), and buffer B is 0.5 M ammonium phosphate (pH 3.4). (B) For reversed-phase chromatography, nucleotide standards were separated on a C18 column by using a method that started with a 70:30 concentration (A:B) at a flow rate of 0.8 ml/min, which was changed in a linear gradient to 40:60 (A:B) after 30 min. From 30 to 60 min the gradient was changed linearly from 40:60 (A:B) to 0:70:30 (A:B:C). Buffer A consisted of 5 mM t-butyl ammonium phosphate (PicA Reagent), 10 KH2PO4, and 0.25% methanol adjusted to pH 6.9. Buffer B consisted of 5 mM t-butyl ammonium phosphate, 50 mM KH2PO4, and 30% methanol (pH 7.0). Buffer C was acetonitrile. The absorbances were measured at 254 nm.

View this table: [in this window] [in a new window]

TABLE 1. Comparison of C18 and SAX columnsa

To further test the peak purity on the columns, selected peaks, including ADP and all of the NTPs, were collected from the SAX column and run on the C₁₈ column and vice versa (data not shown). Every isolated peak from the SAX ran cleanly on C₁₈, with no contaminating peaks with the exception of UTP, which had a small unknown contaminating peak (16%) at 43 min. In contrast, the GTP/UTP peak from the C₁₈ column separated into distinct UTP and GTP peaks on the SAX column, and the ATP peak contained some ppGpp, confirming our observation that there was comigration of these peaks on the C₁₈ column. Selected peaks were also analyzed by MALDI for purity. Ultimately, every sample was run on both columns to optimize separation and quantitation.

Because of the multiple recovery and concentration steps in the method, UV spectrometry and TLC of radioactive nucleotides were done at each step to assess recovery. The recovery from the Q-Sepharose collecting column for di- and triphosphate was greater than 90% (Table 2). Retention of monophosphates, however, was very poor (10%). Recovery from the 100-molecular-weight-cutoff dialysis tubing after 40 h was ca. 80% (Table 2).

View this table: [in this window] [in a new window]

TABLE 2. Analysis of sample loss and recovery with the new methoda

Pool size lability. When cells were initially centrifuged at 37°C followed by cold acid extraction (simply scaling up the method of Bochner and Ames [4]), the pool sizes were not in concert with expectations. The ATP peak, expected to be the most abundant nucleotide in the extract by two- to threefold, was quite small (Fig. 3A). In fact, the nucleotide yield in general, especially the NTPs, was quite low compared to published results. After we attempted numerous modifications in the extraction methodology, such as using vacuum filtration or fixing the cells with formaldehyde prior to centrifugation and doing a basic extraction, it became clear that the result was most likely caused by the manipulation of the cells prior to extraction.

View larger version (14K): [in this window] [in a new window]

FIG. 3. Effect of manipulation prior to extraction on the intracellular nucleotides pools. E. coli MG1655 was grown in 0.4% glucose-MOPS medium to an OD600 of 0.5. At that point, cells were extracted according to different procedures to assess the affect on pools sizes. (A) Cells were centrifuged prior to extraction. The supernatant was discarded, and the pellet was resuspended in 1 M formic acid. The extracted cells were then centrifuged and discarded, leaving behind the supernatant containing the nucleotides. The nucleotides were collected on a Q-Sepharose column, eluted in 1 M ammonium formate, dialyzed, lyophilized, and injected into the HPLC. (B) Cells were treated according to the new method by pouring the cell suspension into 23 M formic acid to a final concentration of 1 M, fast frozen, rapidly thawed, and extracted on ice for 30 min, followed by the method stated above. (C) Prior to the addition of acid, the cells were allowed to rest on the bench top without shaking for 2 min, followed by the same method as for panel B. (D) Prior to the addition of acid, the cells were allowed to rest on the bench top without shaking for 5 min, followed by the same method as panel B. (E) Illustration of the chromatograms showing ADP (blue), ATP (maroon), and ppGpp (white).

To test this hypothesis, we attempted to kill the cells and extract the nucleotides in the quickest, least manipulative way possible by adding formic acid directly to the culture at the desired cell density; thereby, our new methodology was developed. The results indicated that the direct addition of formic acid into the culture medium, followed by column concentration, dialysis, and lyophilization, produced the correct nucleotide ratios as reported in the published literature (Fig. 3B and Table 3). We further assessed the effect of preparation technique on the nucleotide pools by simply letting the culture sit on the bench for 2 or 5 min without any other manipulation prior to acid addition. We found that ATP rapidly dropped almost twofold in 2 min and ppGpp increased fourfold in 5 min (Fig. 3C to E). It is evident that spinning the cells and allowing them to settle results in two different physiologic processes, since the nucleotide profiles are quite different: although the relative amount of ATP is the same, there is considerably more ppGpp in the latter profile.

View this table: [in this window] [in a new window]

TABLE 3. Nucleotide concentrations of E. coli in mid-log phasea

Pool quantitation. After reproducing the predicted ratios for the nucleotides, we were able to calculate the nucleotide concentrations per cell by quantitating the amount of each nucleotide extracted from the mixture (correlated to known standards) and dividing by the cell number and assumed cellular volume. In 100 ml of a culture with an OD₆₀₀ of 0.5, there are approximately 2 x 10¹⁰ cells (assessed by plating). Taking an average volume per cell in logarithmic phase to be approximately 10^–15 liters (6, 10, 18), the concentration of ATP is, at a lower limit, 2.5 mM. If we assume a 71% recovery (the cumulative recovery from the Q-Sepharose and dialysis steps), the concentration of ATP is approximately 3.5 mM. The concentrations of other nucleotides for mid-log cells are included in Table 3 and, in general, correlate well with published data. There are some notable discrepancies between the Bochner and Ames data and the data presented here: the GTP concentration using this technique is almost twofold higher. Acetyl coenzyme A (acetyl-CoA) is approximately sixfold higher by our measurements, and UDP(G) is approximately threefold higher.

Growth rate dependence. The persistence of growth rate dependence even in Rel/SpoT knockouts has lead to numerous theories as to ppGpp-independent causes of this effect. One such theory is that the concentration of the principal iNTP, ATP, changes with growth rate, and this might affect transcription from rRNA promoter P1. Our analyses confirm observations from several groups that ATP and GTP concentration do not significantly change at different growth rates (Fig. 4A) (21, 24). However, our method offers a quantitative advantage. Based on our calculations, the concentration of ATP is above 3 mM for each growth rate in mid-log phase. The method does confirm that the concentration of ppGpp is inversely proportional to the growth rate (Fig. 4B). Our calculations suggest that the concentration of ppGpp is approximately 260, 113, and 63 µM in 0.4% glycerol, 0.4% glucose, and 0.4% glucose and all 20 amino acid MOPS media, respectively.

View larger version (18K): [in this window] [in a new window]

FIG. 4. Intracellular pool sizes of ATP, GTP, and ppGpp with different growth rates. Using the new method, E. coli MG1655 was grown in selected MOPS medium containing 0.4% glycerol, 0.4% glucose, or 0.4% glucose and 80 µg of the 20 amino acids/ml, which grow at approximately 0.5, 1, and 2 divisions per h. At an OD600 of 0.5, cells were immediately mixed with formic acid and treated according the method above. (A) Nanomoles of ATP and GTP extracted at growth rates of 0.5 divisions/h (blue), 1 division/h (maroon), and 2 divisions/h (white). (B) Nanomoles of ppGpp extracted per 109 cells at the growth rates given above.

Growth curve. One advantage of the HPLC method described here is the ability to accurately measure multiple nucleotide species through the growth curve. We tracked the complete growth curve of E. coli MG1655 in 0.4% glucose MOPS medium from an OD₆₀₀ of 0.5 as it goes into early stationary phase. The ATP curve follows a pattern that has been well described in the literature: as the cells enter stationary phase, the amount of ATP decreases between two- and threefold from the starting concentration (Fig. 5A). In our case, it dropped 2-fold. The most dramatic drop (2.5-fold) occurs when the cells first transition from logarithmic growth into stationary phase. This drop is followed by a transient resurgence of ATP ("rebound") and then a decrease as the cultures go deeper into stationary phase. It is interesting that the ATP pools began to decrease while the cells were still in logarithmic phase. Analysis of ADP pools shows a transient increase when the ATP pools begin to decrease; at maximum, there is a fourfold increase while the cells complete exponential phase (Fig. 5B). However, the absolute increase in ADP is insufficient to completely account for the drop in ATP. AMP, as mentioned in Materials and Methods, is not quantified by this method.

View larger version (13K): [in this window] [in a new window]

FIG. 5. Nucleotide pools during the transition from exponential growth into stationary phase. E. coli MG1655 were grown in 0.4% glucose MOPS with 10 µg of thiamine/ml from a 1:200 dilution of a starter culture grown overnight. At different time points starting with an OD600 of 0.5, an adjusted volume of cells was removed from the culture and mixed directly with formic acid, followed by the extraction method recorded above. The volume of culture at each time point was adjusted to correspond to 50 ml at an OD600. The data points were collected until 6 to 8 h into stationary phase. Shown in all graphs is an example of a typical growth curve. Shown are the results of three experiments. (A) Purine pools: ATP (blue), GTP (pink), and ppGpp (yellow); (B) NDP pools: ADP (blue), UDP (pink), and GDP (yellow); (C) pyrimidine pools: UTP (pink) and CTP (blue); (D) dNTP pools: dATP (turquoise), dCTP (blue), dGTP (pink), and dTTP (yellow).

GTP exhibits a pattern through the growth curve very much like ATP, with a precipitous drop as the cells transition from exponential to nonexponential growth, followed by an increase and then a decrease (Fig. 5A). However, the GTP drop is larger than that of ATP (i.e., fourfold), and the "rebound" following the transition is not as large as that of ATP. The drop in GTP corresponds to the increase in ppGpp. Unlike GTP, the drop in GDP as the cells enter into stationary phase is significantly less, and the concentration is stable thereafter. The ppGpp profile looked like a classic stringent response. There is a low basal level in exponential phase, a sharp spike as the cells enter stationary phase, and a slow return to a new baseline level higher than the prestringent levels.

With regard to the pyrimidines, the UTP pools exhibited an interesting behavior contrary to that observed with the purine samples (Fig. 5C). Although they showed the same drop (almost twofold) that was observed with ATP and GTP during the transition from log phase to stationary phase, there was a progressive increase in UTP as the cells moved into late stationary phase. By 15 h, there was a 1.4-fold increase, and it appears the trend was still increasing past 15 h. There is no commensurate decrease in UDP, which appears to remain relatively constant through the growth curve (Fig. 5B).

Of all of the nucleotides, CTP appears to be the most constant through the whole growth curve, with minor fluctuations in its concentration. The decrease in CTP as the cells transition happens 30 min to 1 h after the other nucleotides decrease. There is a small rebound in CTP after the cells enter stationary phase, followed by a decrease. We were unable to measure CDP.

With regard to the dNTPs, dTTP appears to be the most abundant in the cell despite the fact that its precursors, UDP and UTP, are not (Fig. 5D). dATP is the next most abundant, followed by dGTP, which remained quite low in the cell. All deoxynucleotides appeared to decrease through the growth curve ranging from 2.5- to 5-fold. The decrease begins even during balanced growth and does not have the same abrupt drop, as seen with other nucleotides when the cells transition from growth to stationary phase. There do seem to be oscillations in pool concentrations through the entire growth curve.

Acetyl-CoA and UDP(G). After we analyzed the nucleotides in the growth curve, it became apparent that some of the largest peaks on the spectrum that exhibited the most dynamic changes were not included among our standard nucleotides. Selected peaks were collected from several runs, pooled together, and identified by using MALDI analysis.

The first peak to be analyzed came out consistently after the largest triphosphate peaks on both the C₁₈ and SAX column, and it appeared to have the spectral characteristics of adenine. Using MALDI, a peak with molecular weight of 808.69 was detected. Based on mass, the peak was most likely acetyl-CoA, which has a molecular weight of 809.57. The slight difference in mass is attributed to calibration of the mass spectrometer. To confirm the identity of the peak, an acetyl-CoA standard was run that correctly matched the unknown peak. The acetyl-CoA pool exhibits a pattern through the growth curve not unlike many of the other nucleotides (Fig. 6A). Once again, the drop in acetyl-CoA begins in "mid-log" growth. There is a relatively stable baseline concentration which drops almost twofold when the cells first begin to transition into stationary phase. A new baseline value, albeit lower, is achieved.

View larger version (14K): [in this window] [in a new window]

FIG. 6. Additional nucleotide pools during the transition from exponential growth into stationary phase. Several peaks that were not included in the original standards were observed on our chromatograms. The most abundant of these peaks were collected from the SAX column, pooled from different samples, and analyzed by using MALDI. Based on mass, the molecules were acetyl-CoA, UDP(G), and an unknown molecule. The identity of acetyl-CoA and UDP(G) was confirmed by comigration with standards. (A) Acetyl-CoA; (B) UDP-glucose/galactose.

One of the most prominent peaks in the growth curve eluted early (12 min on the SAX column, 10 min on the C₁₈ column), had uridine spectral characteristics, and had a formula mass of 565.7 as determined by MALDI analysis. Using the Bochner and Ames data to make an educated guess as to the nature of the peak, we determined that the peak was most likely UDP-glucose or UDP-galactose (the HPLC is not able to separate these isomers), which has a molecular mass of 566.3. The peak identity was confirmed by comigration with a standard. The UDP(G) peak exhibited a very unique behavior. Its starting concentration by our measurement was 1.5 mM, almost threefold higher than previously reported (Fig. 6B) (4). Like every other nucleotide other than ADP and ppGpp, it too dropped 1.4-fold upon growth transition. However, there was a very rapid increase in UDP(G) just as the cells entered into stationary phase. The rapid increase appears to coincide with the point where the cells began to reduce their rate of division. The concentration of UDP(G) also remains quite high into early stationary phase: 1.2-fold higher than its initial value and 1.7-fold higher than the trough value.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
In the present study, we modified existing approaches for analyzing nucleotide pools in bacteria that, in turn, exposed new information about pool sizes and changes. This method offers several advantages over current approaches. First, it is quantitative and enables the analysis of changes in nucleotide concentrations under a variety of conditions with minimal manipulation of cells prior to extraction. Second, using a two-column approach allowed rapid and excellent separation of all of the nucleotides. Third, since the nucleotides are subjected to HPLC, they are easily collected and analyzed via mass spectrometry. The major deficiency in this method lies primarily in its cumbersome nature, which involves collecting large culture volumes over multiple time points and many preparative steps before the results can be analyzed. The method may also not detect very low concentration compounds. For instance, we were never able to see pppGpp, although it appears on TLC (4). There might be some acid hydrolases that destroy compounds such as the pentaphosphate before the large samples are able to be frozen. Lastly, since the nucleotides are collected on Q-Sepharose, many important molecules such as the uncharged species and monophosphates are lost, leaving unanswered questions about nucleotide conservation and production.

The sensitivity of the nucleotide pools to preparative steps was a striking observation in the present study. This has been suggested for many years; however, to our knowledge, no clear demonstration of this phenomenon has been presented in the literature. The turnover of cytosolic ATP in yeast has been estimated at 1.5 mM s^–1 (7, 12). Therefore, it is not surprisingly that not only centrifugation but even short (2 min) lapses in sample preparation dramatically affect results. This occurs even when the cells are in mid-log phase (OD₆₀₀ of 0.5), so we suggest that the effect is most likely due to hypoxia. Exactly how the hypoxia translates into a stringent response is not entirely clear. This observation requires a careful reevaluation of many measurements in the literature since many reported concentrations of ATP, NAD/NADH ratios, ppGpp, etc., come from samples that have undergone multiple preparative manipulations prior to extraction such as washing in phosphate-buffered saline or centrifugation.

Quantitations from the new method closely reflect those obtained in previous analyses (4). Aside from the obvious difference that Bochner and Ames studied Salmonella enterica serovar Typhimurium and we studied E. coli, there are some notable differences in our data such as higher concentrations of GTP, UDP(G), and acetyl-CoA. Using HPLC, we were able to measure actual nanomoles of material collected for a given number of cells; Bochner and Ames, in contrast, assumed an ATP concentration (from an unreferenced previous measurement) and quantitated other nucleotides based on spot intensity and the number of phosphates. Also, using a dual-column approach, we obtained excellent separations, perhaps with less overlap than a TLC approach. Bochner and Ames also reported that they took measurements at an OD₆₅₀ between 0.5 to 1.0 from a culture with a maximum OD₆₅₀ of 1.2. Based on our analysis of the growth curve, most nucleotides already begin to drop at the end of logarithmic phase even though the cells are still dividing. Our measurements were all taken at an OD₆₀₀ of 0.5, and our cultures did not reach stationary phase until an OD₆₀₀ of 2.0, better capturing exponential growth concentrations.

It should be noted one of the biggest difficulties in reporting true concentrations is determining the cell volume and CFU/ml of culture used to estimate the concentration. We decided to approximate a cell volume of 1 µm³ based on growth rate, but volume estimates range from 0.6 to 1.5 (6, 10, 18). Also, although we determined the CFU/ml at an OD₆₀₀ of 0.5 in 0.4% glucose-MOPS to be between 2 x 10⁸ and 3 x 10⁸ cells/ml, others report CFU/ml values of up to 7 x 10⁸ to 8 x 10⁸ cells/ml (6). Thus, the concentrations we report, although unlikely to change dramatically, are limited by the variability in these measurements. We did not even attempt to convert data from later points in the growth curve into actual concentrations since the cell size here can be extremely variable (1, 18).

Our observations confirm that which seems to be universally accepted: that the ATP pools do not change with growth rate. Using HPLC, it appears that the concentration at all three growth rates is at least 3 mM. Although this concentration is well above the 0.8 mM needed to saturate the rRNA promoter P1 in vitro, in vivo transcription, with its multiple intracellular components and topological factors, might not be saturated at this concentration of ATP (15). In fact, evidence suggests that rRNA promoters are sensitive to changes in ATP concentrations in vivo (27). Schneider and Gourse have also astutely noted that any extraction technique does not necessarily give the "free and available" pool size, which might be much less than what is extracted (28). Our data do not directly address this problem.

One of the most interesting results to come from the present study is the analysis of the nucleotide pools as the cells transition from growth to stationary phase. We observed that many of the pools, including ATP, GTP, and UTP begin to decrease while the cells are still in mid-log phase. This might reflect that although the cells are still dividing at a maximal rate, nutrients are starting to be limiting. It could be that once nutrient availability decreases past a critical threshold, a stringent response is elicited, significantly reducing cell division. This would be reflected in the precipitous drop in most nucleotides right at the induction of the stationary phase. Truly balanced-growth is a very ephemeral state and, as our data demonstrate, may only be present at optical densities lower than 0.5 (22).

The second major observation from our growth curve analysis is that after an almost universal drop in nucleotide concentration as the cells transition from growth to stationary phase, there is a "rebound" of certain nucleotides, most notably ATP, after the cells enter stationary phase. It was also observed for the other NTPs and ADP, although to lesser extents. Others have observed a similar "rebound" in ATP after a stringent response (14). The cause of the bounce is unclear, but we hypothesize that it might reflect a disconnect between the cell's metabolic machinery which is geared for maximal growth and a sudden reduction in the rate of cell division. It appears to take approximately 1.5 h for the cells to readjust their metabolism to slower growth. Of note, our results, as well as those of Fiil and von Meyenburg, were only attempted in minimal media and might be a function of medium-dependent growth limitation, as evidenced by the sharpness of our OD₆₀₀ curve (14). It would be interesting to repeat the analysis as cells enter stationary phase in a complete medium.

Another interesting observation from the growth curve analysis is that there is an increase in UTP into stationary phase despite a general decline in the other nucleotides. The rise in UTP might reflect a cellular preservation of the pyrimidine pathways since CTP does not appear to decrease either. The higher UTP values might also be related to the elevated UDP-glucose/galactose. This result was particularly surprising in lieu of evidence suggesting that UDP-glucose inhibits ^s, a major transcription factor involved in the stationary phase (5). However, that study never directly measured the UDP(G) pools, and the authors only observed increases in ^s in a galU mutant under growth conditions, a state which might not reflect ^S physiology for the stationary phase. Our data are confirmed by another study that observed increased levels of UDP(G) under slow-growth conditions (31). The rise in UDP(G) might have some other signaling capability either directly or through its downstream products trehalose and membrane-derived oligosaccharides.

The concentrations of acetyl-CoA reported here were surprising for two reasons. First, our mid-log concentration was 6-fold higher than that detected in the Bochner and Ames study. We believe our measurement to be very accurate since we assessed peak purity via MALDI and cannot explain why their value is so much lower. Although not directly reported, we estimated from a recent report studying acetyl phosphate in E. coli that the log-phase acetyl-CoA concentration is at least 600 µM, assuming all three phosphates in acetyl-CoA turn over as rapidly as the phosphates in ATP (17). This result appears to be closer to our measurement than to that of Bochner and Ames, and the twofold difference from our result might reflect different strains and the effect of pyruvate supplementation. Second, even after stationary phase, the acetyl-CoA reached a new basal value that is in contrast to reports that acetyl-CoA can drop to within 5% of its starting levels in stationary phase (30). However, that study only measured acetyl-CoA indirectly, and the authors used a glucose-deprived culture, whereas we were able to measure acetyl-CoA directly.

 
* Corresponding author. Mailing address: Division of Infectious Diseases, Department of Medicine, 522 Johnson Pavilion, 36th and Hamilton Walk, University of Pennsylvania, Philadelphia, PA 19104. Phone: (215) 662-6475. Fax: (215) 573-1240. E-mail: rubinh{at}mail.med.upenn.edu

Published ahead of print on 26 October 2007.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1
1. Akerlund, T., K. Nordstrom, and R. Bernander. 1995. Analysis of cell size and DNA content in exponentially growing and stationary-phase batch cultures of Escherichia coli. J. Bacteriol. 177:6791-6797.[Abstract/Free Full Text]
2
2. Alper, S., L. Duncan, and R. Losick. 1994. An adenosine nucleotide switch controlling the activity of a cell type-specific transcription factor in Bacillus subtilis. Cell 77:195-205.[CrossRef][Medline]
3
3. Avarbock, D., A. Avarbock, and H. Rubin. 2000. Differential regulation of opposing RelMtb activities by the aminoacylation state of a tRNA. Ribosome, mRNA, RelMtb complex. Biochemistry 39:11640-11648.[CrossRef][Medline]
4
4. Bochner, B. R., and B. N. Ames. 1982. Complete analysis of cellular nucleotides by two dimensional thin-layer chromatography. J. Biol. Chem. 257:9759-9769.[Abstract/Free Full Text]
5
5. Bohringer, J., D. Fischer, G. Mosler, and R. Hengge-Aronis. 1995. UDP-glucose is a potential intracellular signal molecule in the control of expression of sigma-s and sigma-s dependent genes in Escherichia coli. J. Bacteriol. 177:413-422.[Abstract/Free Full Text]
6
6. Bremer, H., and P. P. Dennis. 1996. Modulation of chemical composition and other parameters of the cell by growth rate, p. 1553-1569. In F. C. Niedhardt, R. Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, 2nd ed. ASM Press, Washington, DC.
7
7. Bucholz, A., R. Takors, and C. Wandrey. 2001. Quantification of intracellular metabolites in Escherichia coli K-12 using liquid chromatographic-electrospray ionization tandem mass spectrometric techniques. Anal. Biochem. 295:129-137.[CrossRef][Medline]
8
8. Cashel, M., D. R. Gentry, V. J. Hernandez, and D. Vinella. 1996. The stringent response, p. 1458-1496. In F. C. Niedhardt, R. Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, 2nd ed. ASM Press, Washington, DC.
9
9. Chang, D., D. J. Smalley, and T. Conway. 2002. Gene expression profiling of Escherichia coli growth transitions: an expanded stringent response model. Mol. Microbiol. 45:289-306.[CrossRef][Medline]
10
10. Cox, R. A. 2004. Quantitative relationships for specific growth rates and macromolecular compositions of Mycobacterium tuberculosis, Streptomyces coelicolor A3(2) and Escherichia coli B/r: an integrative theoretical approach. Microbiology 150:1413-1426.[Abstract/Free Full Text]
11
11. Dahl, J. L., C. N. Kraus, H. I. Boshoff, B. Doan, K. Foley, D. Avarbock, G. Kaplan, V. Mizrahi, H. Rubin, and C. E. Barry III. 2003. The role of RelMtb-mediated adaptation to stationary phase in long-term persistence of Mycobacterium tuberculosis in mice. Proc. Natl. Acad. Sci. USA 100:10026-10031.[Abstract/Free Full Text]
12
12. De Koning, W., and K. van Dam. 1992. A method for the determination of changes of glycolytic metabolites in yeast on a subsecond time scale using extraction at neutral pH. Anal. Biochem. 204:118-123.[CrossRef][Medline]
13
13. Dennis, P. P., M. Ehrenberg, and H. Bremer. 2004. Control of rRNA synthesis in Escherichia coli: a systems biology approach. Microbiol. Mol. Biol. Rev. 68:639-668.[Abstract/Free Full Text]
14
14. Fiil, N. P., and K. von Meyenburg. 1972. Accumulation and turnover of guanosine tetraphosphate in Escherichia coli. J. Mol. Biol. 71:769-783.[CrossRef][Medline]
15
15. Gaal, T., M. S. Bartlett, W. Ross, C. L. Turnbough, Jr., and R. L. Gourse. 1997. Transcription regulation by initiating NTP concentration: rRNA synthesis in bacteria. Science 278:2092-2097.[Abstract/Free Full Text]
16
16. Huang, D., Y. Zhang, and X. Chen. 2003. Analysis of intracellular nucleoside triphosphate levels in normal and tumor cell lines by high-performance liquid chromatography. J. Chromatogr. B Anal. Technol. Biomed. Life. Sci. 782:101-109.
17
17. Klein, A. H., A. Shulla, S. A. Reimann, D. H. Keating, and A. J. Wolfe. 2007. The intracellular concentration of acetyl phosphate in Escherichia coli is sufficient for direct phosphorylation of two-component response regulators. J. Bacteriol. 189:5574-5581.[Abstract/Free Full Text]
18
18. Kubitschek, H. E., and J. A. Friske. 1986. Determination of bacterial cell volume with the Coulter Counter. J. Bacteriol. 168:1466-1467.[Abstract/Free Full Text]
19
19. Little, R., and H. Bremer. 1982. Quantitation of guanosine 5',3'-bisdiphosphate in extracts from bacterial cells by ion-pair reverse-phase high-performance liquid chromatography. Anal. Biochem. 126:381-388.[CrossRef][Medline]
20
20. Mitani, T., J. Heinze, and E. Freese. 1977. Induction of sporulation in Bacillus subtilis by decoyinine or hadacidin. Biochim. Biophys. Acta 77:1118-1125.
21
21. Murray, H. D., D. A. Schneider, and R. L. Gourse. 2003. Control of rRNA expression by small molecules is dynamic and nonredundant. Mol. Cell 12:125-134.[CrossRef][Medline]
22
22. Neidhardt, F. C. 1999. Bacterial growth: constant obsession with dN/dt. J. Bacteriol. 181:7405-7408.[Free Full Text]
23
23. Ochi, K., J. Kandala, and E. Freese. 1982. Evidence that Bacillus subtilis sporulation induced by the stringent response is caused by the decrease in GTP or GDP. J. Bacteriol. 151:1062-1065.[Abstract/Free Full Text]
24
24. Paul, B. J., W. Ross, T. Gaal, and R. L. Gourse. 2004. rRNA transcription in Escherichia coli. Annu. Rev. Genet. 38:749-770.[CrossRef][Medline]
25
25. Payne, S. M., and B. N. Ames. 1982. A procedure for rapid extraction and high-pressure liquid chromatographic separation of the nucleotides and other small molecules from bacterial cells. Anal. Biochem. 123:151-161.[CrossRef][Medline]
26
26. Ryals, J., R. Little, and H. Bremer. 1982. Control of rRNA and tRNA synthesis in Escherichia coli by guanosine tetraphosphate. J. Bacteriol. 151:1261-1268.[Abstract/Free Full Text]
27
27. Schneider, D. A., T. Gaal, and R. L. Gourse. 2002. NTP-sensing by rRNA promoters in Escherichia coli is direct. Proc. Natl. Acad. Sci. USA 99:8602-8607.[Abstract/Free Full Text]
28
28. Schneider, D. A., and R. L. Gourse. 2004. Relationship between growth rate and ATP concentration in Escherichia coli. J. Biol. Chem. 279:8262-8268.[Abstract/Free Full Text]
29
29. Sonenshein, A. L. 2005. CodY: a global regulator of stationary phase and virulence in gram-positive bacteria. Curr. Opin. Microbiol. 8:203-207.[CrossRef][Medline]
30
30. Takamura, Y., and G. Nomura. 1988. Changes in the intracellular concentration of acetyl CoA and malonyl-CoA in relation to the carbon and energy metabolism of Escherichia coli K-12. J. Gen. Microbiol. 134:2249-2253.[Abstract/Free Full Text]
31
31. Tweendale, H., L. Notley-McRobb, and T. Ferenci. 1998. Effect of slow-growth on metabolism of Escherichia coli, as revealed by global metabolite pool ("metabolome") analysis. J. Bacteriol. 180:5109-5116.[Abstract/Free Full Text]
32
32. Walker, K. A., P. Mallik, T. S. Pratt, and R. Osuna. 2004. The Escherichia coli fis promoter is regulated by changes in the levels of its transcription initiation nucleotide CTP. J. Biol. Chem. 279:50818-50828.[Abstract/Free Full Text]

---

Journal of Bacteriology, January 2008, p. 718-726, Vol. 190, No. 2
0021-9193/08/$08.00+0     doi:10.1128/JB.01020-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

This article has been cited by other articles:

• Lee, S. J., Trostel, A., Le, P., Harinarayanan, R., FitzGerald, P. C., Adhya, S. (2009). Cellular stress created by intermediary metabolite imbalances. Proc. Natl. Acad. Sci. USA 106: 19515-19520 [Abstract] [Full Text]  
• Hirayama, A., Kami, K., Sugimoto, M., Sugawara, M., Toki, N., Onozuka, H., Kinoshita, T., Saito, N., Ochiai, A., Tomita, M., Esumi, H., Soga, T. (2009). Quantitative Metabolome Profiling of Colon and Stomach Cancer Microenvironment by Capillary Electrophoresis Time-of-Flight Mass Spectrometry. Cancer Res. 69: 4918-4925 [Abstract] [Full Text]  
• Behrends, V., Ebbels, T. M. D., Williams, H. D., Bundy, J. G. (2009). Time-Resolved Metabolic Footprinting for Nonlinear Modeling of Bacterial Substrate Utilization. Appl. Environ. Microbiol. 75: 2453-2463 [Abstract] [Full Text]  
• Nakane, S., Nakagawa, N., Kuramitsu, S., Masui, R. (2009). Characterization of DNA polymerase X from Thermus thermophilus HB8 reveals the POLXc and PHP domains are both required for 3'-5' exonuclease activity. Nucleic Acids Res 37: 2037-2052 [Abstract] [Full Text]  
• Toth, M., Chow, J. W., Mobashery, S., Vakulenko, S. B. (2009). Source of Phosphate in the Enzymic Reaction as a Point of Distinction among Aminoglycoside 2''-Phosphotransferases. J. Biol. Chem. 284: 6690-6696 [Abstract] [Full Text]  
• Mowa, M. B., Warner, D. F., Kaplan, G., Kana, B. D., Mizrahi, V. (2009). Function and Regulation of Class I Ribonucleotide Reductase-Encoding Genes in Mycobacteria. J. Bacteriol. 191: 985-995 [Abstract] [Full Text]  
• Rofougaran, R., Crona, M., Vodnala, M., Sjoberg, B.-M., Hofer, A. (2008). Oligomerization Status Directs Overall Activity Regulation of the Escherichia coli Class Ia Ribonucleotide Reductase. J. Biol. Chem. 283: 35310-35318 [Abstract] [Full Text]  
• Haddix, P. L., Jones, S., Patel, P., Burnham, S., Knights, K., Powell, J. N., LaForm, A. (2008). Kinetic Analysis of Growth Rate, ATP, and Pigmentation Suggests an Energy-Spilling Function for the Pigment Prodigiosin of Serratia marcescens. J. Bacteriol. 190: 7453-7463 [Abstract] [Full Text]  
• Noriega, C. E., Schmidt, R., Gray, M. J., Chen, L.-L., Stewart, V. (2008). Autophosphorylation and Dephosphorylation by Soluble Forms of the Nitrate-Responsive Sensors NarX and NarQ from Escherichia coli K-12. J. Bacteriol. 190: 3869-3876 [Abstract] [Full Text]  

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Buckstein, M. H. Articles by Rubin, H. Search for Related Content PubMed PubMed Citation Articles by Buckstein, M. H. Articles by Rubin, H.