Genetic response to metabolic fluctuations: correlation between central carbon metabolism and DNA replication in Escherichia coli

Background Until now, the direct link between central carbon metabolism and DNA replication has been demonstrated only in Bacillus. subtilis. Therefore, we asked if this is a specific phenomenon, characteristic for this bacterium and perhaps for its close relatives, or a more general biological rule. Results We found that temperature-sensitivity of mutants in particular genes coding for replication proteins could be suppressed by deletions of certain genes coding for enzymes of the central carbon metabolism. Namely, the effects of dnaA46(ts) mutation could be suppressed by dysfunction of pta or ackA, effects of dnaB(ts) by dysfunction of pgi or pta, effects of dnaE486(ts) by dysfunction of tktB, effects of dnaG(ts) by dysfunction of gpmA, pta or ackA, and effects of dnaN159(ts) by dysfunction of pta or ackA. The observed suppression effects were not caused by a decrease in bacterial growth rate. Conclusions The genetic correlation exists between central carbon metabolism and DNA replication in the model Gram-negative bacterium, E. coli. This link exists at the steps of initiation and elongation of DNA replication, indicating the important global correlation between metabolic status of the cell and the events leading to cell reproduction.


Background
When considering a bacterial cell as a microbial factory, producing various macromolecules either natural or formed due to introduction of recombinant genes, several biochemical processes must be taken into consideration. Among them, there are two basic processes ensuring that more specialized reactions (like transcription of particular genes and translation of particular mRNAs on ribosomes as well as enzyme-mediated production of various compounds) can occur. These two processes are central carbon metabolism (for a review see ref. [1]) and DNA replication (for a review see ref. [2]). The former one provides energy from nutrients, which is absolutely necessary to all life functions of cells. The latter one, although consuming cellular energy, ensures integrity of genetic material and its inheritance by daughter cells after each cell division, providing the source of information about biological structures and functions of macromolecules.
The central carbon metabolism (CCM) is generally recognized as a set of biochemical pathways devoted to transport and oxidation of main carbon sources in the cell [1]. In a model Gram-negative bacterium, Escherichia coli, it consists of the phosphortransferase system, glycolysis, gluconeogenesis, pentose-monophosphate bypass with Entner-Dudoroff pathway, Krebs cycle with glyoxylate bypass and the respiration chain [3]. Biochemical reactions of these pathways ensure the optimal energy production and usage in the cell at particular growth conditions, in order to keep homeostasis.
DNA replication is a process of genetic information duplication, which is necessary to equal and precise distribution of the genetic material to both daughter cells after each cell division [2]. The process of replicative DNA synthesis requires large cellular machinery, which in E. coli consists of DNA polymerase III holoenzyme (containing at least 10 subunits) and other essential proteins, including DnaB helicase and DnaG primase.
Although it was observed previously that regulation of DNA replication may depend on bacterial cell metabolism, it was generally assumed that this dependency is indirect. For example, it might result from different availability of cellular energy and/or precursors of macromolecules [5,6] or from production of specific alarmons, like cyclic AMP (cAMP) [7,8] or guanosine tetraphosphate (ppGpp) [9][10][11][12], in response to nutritional deprivations. However, it was reported recently that DNA replication may be directly linked to central carbon metabolism, particularly glycolysis, in a model Gram-positive bacterium, Bacillus subtilis [13]. Namely, specific suppression of conditionally-lethal (temperaturesensitive, ts) mutations in genes coding for replication proteins (DnaE, a DNA polymerase involved in lagging strand synthesis, DnaC, a helicase -homologue of E. coli DnaB protein, and DnaG, the primase) by dysfunction of certain genes coding for enzymes involved in glycolysis, was observed. An indirect suppression mechanism (e.g. by slowing down of bacterial growth rate) was excluded, strongly suggesting a real link between glycolysis and DNA replication. Thus, the existence of such a link should be considered in any studies on both these processes, as well as when constructing and using biotechnological systems for efficient production of desired compounds.
Until now, the direct link between central carbon metabolism and DNA replication has been demonstrated only in B. subtilis [13]. Therefore, we asked if this is a specific phenomenon, characteristic for this bacterium and perhaps for its close relatives, or a more general biological rule. Since E. coli is both a model Gram-negative bacterium and a widely used host for production of recombinant proteins, in our studies, which were performed to answer the above question, we employed strains of this species.

Methods
Bacterial strains, plasmids and bacteriophages E. coli strains used in this work are listed in Table 1. Plasmids and bacteriophages are described in Table 2. New bacterial strains and plasmids were constructed according to standard procedures of P1 transduction and molecular cloning, respectively [14].

Oligonucleotides
Oligunucleotides are described in Table 3.

Growth conditions
Luria -Bertani (LB) medium, and minimal media M9 and MM, were used [14]. Solid media contained 1.5% of bacteriological agar. For liquid cultures, bacteria were grown in various media in shake flasks, with aeration (by shaking). Overnight cultures were diluted in LB and grown to OD 600 = 0.3. Then, 100 μl of the culture or its dilution was plated on solid media. The plates were then incubated at indicated temperatures for indicated time. CFU (colony forming units) were calculated from plates where colony number was between 100 and 1000.

Results
We have employed six E. coli temperature-sensitive mutants in following genes coding for proteins necessary for chromosomal DNA replication: dnaA (coding for the replication initiator protein that binds to the oriC region and forms a specific nucleoprotein structure; this is the first step in the DNA replication initiation), dnaB (coding for the main DNA helicase, the enzyme necessary to melt DNA during the replication process), dnaC (coding for the protein which delivers DnaB helicase to the DnaA protein bound to oriC), dnaE (coding for the α subunit of DNA polymerase III, the catalytic subunit of this enzyme), dnaG (coding for primase, an enzyme necessary to synthesize RNA primers during DNA replication) and dnaN (coding for the β subunit of DNA polymerase III, a protein forming the sliding clamp and allowing DNA polymerase III to be kept on the template DNA strand when synthesizing new polynucleotide strand) [for more detailed information on these genes and their products, see ref. 2]. These mutants are described in Table 1.
To test whether mutations (particularly deletioninsertion mutations) in genes coding for enzymes from central carbon metabolism (CCM) may suppress temperature sensitivity of the replication mutants, we have determined the sensitivity profiles of all tested conditionally lethal mutants. This was necessary to chose temperatures that severely restricted growth of mutant cells, however, which still allowed observing some viability of tested strains; otherwise detection of any suppression would be impossible, as observed in the B. subtilis study [13]. The profiles of temperature-sensitivity of dnaA, dnaB, dnaC, dnaE, dnaG and dnaN mutants in LB medium are shown in Figure 1.
We found no specific suppression (i.e. suppression which could be reversed by plasmid-mediated expression of the wild-type CCM gene whose defective allele resulted in temperature-tolerance of otherwise temperature-sensitive replication mutant) of the effects of dnaC (ts) mutation by any tested dysfunction in the CCM genes ( Figure 3). However, interestingly, efficiency of plating of dnaA46(ts), dnaB8(ts), dnaE486(ts), dnaG(ts) and dnaN159(ts) mutants could be increased by at least one order of magnitude (often considerably more) at sublethal temperatures in the presence of particular mutations in genes coding for enzymes from CCM ( Figure 3). The effects of dnaA46(ts) mutation could be suppressed by dysfunction of pta or ackA, effects of dnaB8(ts) by dysfunction of pgi or pta, effects of dnaE486(ts) by dysfunction of tktB, effects of dnaG(ts) by dysfunction of gpmA, pta or ackA, and effects of dnaN159(ts) by dysfunction of pta or ackA. Most of the suppression phenomena were not complete, i.e. the efficiency of survival of the ts mutants in the sublethal temperature was between 1 and 10% of that in the permissive temperature, though still it was 10 to 100 times higher than that of the ts mutant without suppressor mutation at the sublethal temperature ( Figure 4).  This correlates with the previous findings on the B. subtilis model [13]. Interestingly, the only exceptions were dnaA46 suppressors, restoring 100% growth relative to that under permissive conditions. It is worth noting that dnaA mutants of B. subtilis were not tested in the previous work, mentioned above [13].
To test whether suppressions depicted in Figure 3 were specific, plasmids bearing wild-type copies of disrupted metabolic genes ( Table 2) have been introduced into cells of the double mutants. The wild-type alleles were under control of the pBAD promoter, which could be stimulated by addition of L-arabinose into growth medium. We found that for dnaA46(ts), dnaB8 (ts), dnaE486(ts), dnaG(ts) and dnaN159(ts) hosts, expression of appropriate wild-type allele of CCM gene reversed effects of temperature sensitivity suppression by the corresponding mutant allele (Figure 4). Therefore, we conclude that the suppression effects depicted in Figure 3 are specific for certain mutations.
We asked whether the suppression of temperature sensitivity of mutants in the replication genes by dysfunction of particular genes coding for CCM enzymes could be caused by decreased growth rates of double mutants. This question was substantiated by the fact that DNA replication regulation is known to be dependent on bacterial growth rate [2]. However, we found that although in most cases (excluding the dnaA46 mutants) at 30°C the growth rates of the double mutants revealing suppression of the temperature sensitivity were lower than in wild-type bacteria, a similar or lower decrease in the growth rate was observed also in double mutants which did not suppress the temperature sensitivity ( Figure 5). Therefore, we conclude that the observed suppression effects could not be caused simply by a decrease in bacterial growth rate.
We have also tested whether the suppression can be caused by growth of the replication mutants in media containing various carbon sources, which also allow for different growth rates. Therefore, we have plated dnaA46(ts), dnaB8(ts), dnaC(ts), dnaE486(ts), dnaG(ts) and dnaN159(ts) mutants on plates containing a minimal medium supplemented with various carbon sources: glucose, glycerol, maleic acid or sodium acetate. However, in these experiments, we did not observe any improvement in viability of these mutants at the sublethal temperatures (data not shown). These results corroborate the results of experiments with growth rate measurement, and support our conclusion that the suppression of temperature sensitivity of the replication mutants cannot be explained by lower growth rates of bacteria.

Discussion
The approach to understand cellular processes as a network of complex relations becomes more appreciated  only nowadays. Two major processes responsible for maintenance and reproduction of the cell (i.e. energy metabolism and DNA replication) were studied mostly independently until recently. A direct link between DNA replication and central carbon metabolism (CCM) has been demonstrated solely for one species of Grampositive bacterium, B. subtilis [13]. This finding was a breakthrough in considering these processes as interrelated. Thus, it was crucial to address the question whether such a phenomenon occurs only in the specific strain or it is more general. Here we present evidence that such a link exists also in E. coli, a model Gramnegative bacterium. Despite the general similarity, there are important differences between suppression of effects of mutations in replication genes by dysfunction of genes coding for enzymes of CCM in E. coli and B. subtilis. According to previous report [13], in B. subtilis, the temperature-sensitivity suppression was detected for only three genes: dnaE, dnaC (an equivalent of the E. coli dnaB gene, coding for helicase) and dnaG. Temperature-sensitive mutants in these genes could grow at elevated temperatures in the presence of additional mutations in gapA, pgk, pgm, eno or pykA. These five genes code for enzymes acting at the late stages of glycolysis and gluconeogenesis. In E. coli, we were able to observe suppression of effects of temperature-sensitive mutations not only in dnaE, dnaB and dnaG genes (like in B. subtilis), but also in dnaN and -perhaps the most surprisinglyin dnaA. Moreover, growth at sublethal temperatures of these mutants was observed under conditions of a lack of enzymes involved not only in glycolysis and gluconeogenesis (pgi and gpmA), but also in other regimens of CCM, namely the pentose phosphate pathway (tktB gene) and the overflow pathway (pta and ackA genes). This suggests that in E. coli the link between DNA replication and CCM may be broader than in B. subtilis. Alternatively, the observed differences might result from a partial exploration of a complex system (only some replication and metabolic genes were tested due to technical reasons, namely unavailability of viable mutants).
For B. subtilis, the target of the regulation by metabolic-related signals was shown to be mostly the elongation of the DNA replication process, though some suppressed replication mutations affected also replication initiation [13]. In E. coli, the evidence presented here shows the link between CCM and replication elongation (represented by enzymes involved in the replication complex), and initiation. One of indispensable regulators of the latter process in E. coli is DnaA protein [15,4]. Thus, the finding of the suppression of dnaA46(ts) conditionally-lethal phenotype by mutants in genes involved in CCM suggests the presence of as yet unidentified correlation. Moreover, the observed suppression was complete (100% survival at sublethal temperature relative to survival at permissive temperature), contrary to those noted for other mutants in replication genes. Both suppressors of the dnaA46(ts) phenotype map in the overflow pathway of CCM. This and the presence of the suppressors in genes of enzymes from other pathways beside glycolysis in E. coli could be explained by (i) partial exploration of the coupling system, (ii) the differences in the replication complexes in E. coli and B. subtilis, and/or (iii) different lifestyles and nutrient requirements of these bacterial species. E. coli, during its life cycle, may be exposed to the abrupt changes in the nutrient availability (the "feastfamine" scenario), which requires a more strict regulation, linking energy turnover and DNA replication, thus,    it may profit from more metabolic sensors. Similarly to B. subtilis, the suppression observed in E. coli was not caused by a decrease in the growth rate. Moreover, the increase in the doubling time of replication mutants (by growth on the minimal media containing various carbon sources, including very poor ones, like maleic acid or acetate) did not improve their viability at sublethal temperatures.

CFU/ml
The proposed mechanism of the regulation of DNA replication by CCM in B. subtilis involves a putative metabolic linker which can cause conformational changes in replication proteins to modulate replisome properties [13]. This hypothesis may be supported by the role of acetyl phosphate which can accumulate in the overflow pathway mutants. Acetyl phosphate has been proposed to function as a global signal that fits into various two-compound systems [16,17]. This may require the second, as yet unknown, protein modulating replication proteins, or the mechanism can rely on autophosphorylation. The role of acetyl phosphate in protein folding and stability has been proposed as well [18]. In this light it is interesting that AckA and Pta reduce the production of double-stranded breaks in DNA [19]. Moreover, DiaA, a DnaA-binding protein, contains a SIS motif that might bind phosphosugars [20]. These facts may provide a start point to further works on understanding the link between CCM and DNA replication.
It is worth noting that since we have used deletioninsertion mutants in genes coding for CCM enzymes, the suppressions of the temperature-sensitivity phenotypes of the replication mutants cannot be explained by direct protein-protein interactions. Indeed, numerous and large-scale interactions between replication proteins and CCM enzymes seemed unlikely, which led us to use a set of deletion mutants in tested genes. On the other hand, the use of such mutants ensured that particular enzymatic functions were absent in mutant cells, which excluded potential problems with putative partial inactivation of CCM enzymes caused by point mutations.
One should also take into consideration a possibility that changes in chemical composition of the cells caused by a lack of particular CCM enzymes might alleviate temperature sensitivity of mutants in genes coding for replication proteins. In fact, we cannot exclude that increased concentrations of some substances that accumulate due to metabolic blocks at certain steps of CCM might stabilize the temperature-sensitive replication proteins and allow them to function at higher temperatures. If so, CCM could have no effects on wild-type replication proteins and the DNA replication process in wildtype cells. However, to accept such a hypothesis it would be necessary to assume that there are at least several compounds (metabolites) able to interact specifically with several different temperature-sensitive variants of the replication proteins, resulting in their stabilization at elevated temperatures. Although still possible, such a scenario seems unlikely, therefore, we prefer the hypothesis that there is a link between CCM and DNA replication in bacterial cells.

Conclusions
We show the genetic correlation between central carbon metabolism and DNA replication in the model Gramnegative bacterium, E. coli. Therefore, one might suggest Figure 5 Generation times of double mutants in replication and CCM genes. Bacteria were grown at 30°C in LB and doubling time (values presented in the boxes ± SD) was assessed in the exponential growth phase. The doubling time for the wild-type strain (MG1655) was 48 ± 0.7 min. The colors represent genotypes in which suppressions were observed at sublethal temperatures (red -full suppression, yellow -incomplete suppression). Dash -the generation time was not determined.
that the existence of such a link is a general phenomenon rather than an event occurring very specifically in a small group of organisms. This link exists at the steps of initiation and elongation of DNA replication, indicating the important global correlation between metabolic status of the cell and the events leading to cell reproduction.