Deletion of the 2-acyl-glycerophosphoethanolamine cycle improve glucose metabolism in Escherichia coli strains employed for overproduction of aromatic compounds
© Aguilar et al. 2015
Received: 11 August 2015
Accepted: 11 November 2015
Published: 1 December 2015
As a metabolic engineering tool, an adaptive laboratory evolution (ALE) experiment was performed to increase the specific growth rate (µ) in an Escherichia coli strain lacking PTS, originally engineered to increase the availability of intracellular phosphoenolpyruvate and redirect to the aromatic biosynthesis pathway. As result, several evolved strains increased their growth fitness on glucose as the only carbon source. Two of these clones isolated at 120 and 200 h during the experiment, increased their μ by 338 and 373 %, respectively, compared to the predecessor PB11 strain. The genome sequence and analysis of the genetic changes of these two strains (PB12 and PB13) allowed for the identification of a novel strategy to enhance carbon utilization to overcome the absence of the major glucose transport system.
Genome sequencing data of evolved strains revealed the deletion of chromosomal region of 10,328 pb and two punctual non-synonymous mutations in the dhaM and glpT genes, which occurred prior to their divergence during the early stages of the evolutionary process. Deleted genes related to increased fitness in the evolved strains are rppH, aas, lplT and galR. Furthermore, the loss of mutH, which was also lost during the deletion event, caused a 200-fold increase in the mutation rate.
During the ALE experiment, both PB12 and PB13 strains lost the galR and rppH genes, allowing the utilization of an alternative glucose transport system and allowed enhanced mRNA half-life of many genes involved in the glycolytic pathway resulting in an increment in the μ of these derivatives. Finally, we demonstrated the deletion of the aas-lplT operon, which codes for the main components of the phosphatidylethanolamine turnover metabolism increased the further fitness and glucose uptake in these evolved strains by stimulating the phospholipid degradation pathway. This is an alternative mechanism to its regeneration from 2-acyl-glycerophosphoethanolamine, whose utilization improved carbon metabolism likely by the elimination of a futile cycle under certain metabolic conditions. The origin and widespread occurrence of a mutated population during the ALE indicates a strong stress condition present in strains lacking PTS and the plasticity of this bacterium that allows it to overcome hostile conditions.
KeywordsMetabolic engineering Metabolic plasticity PTS system Escherichia coli Adaptive laboratory evolution 2-Acyl-glycerophosphoethanolamine cycle Glycerol metabolism Glucose metabolism
Adaptation based on mutations is essential to bacterial survival during environmental challenges, such as exposure to antibiotics, variations in oxygen levels or temperature, and during limited availability of carbon [1–4]. Bacterial adaptation that makes living cells extraordinarily plastic is quick and evident within shorts periods of time [1–3], where often beneficial mutations occur that confer increased fitness in new conditions.
These two evolved strains have been studied extensively. Among their main characteristics are that glucose uptake is performed by the GalP permease and phosphorylated by the Glk glucokinase. Additionally, they have an increased glycolytic flux compared to the parental PB11 strain and the co-utilization capacity of several carbon sources because of the absence of catabolic repression due to the lack of the PTS system, as previously described [6, 8, 12, 13]. In addition, whole genome sequence analyses of the PB12 strain allowed for the identification of 23 non-synonymous and 22 synonymous point mutations, as well as a 10,328 bp chromosomal deletion. The effects of some of these genetic changes on the growth of the PB12 strain were also reported .
To determine if both the PB12 and PB13 evolved strains followed similar or independent adaptive paths and obtain a deeper insight into the genetic changes that these evolved strains developed to overcome the absence of the major glucose transport system, in this contribution whole genome sequencing of PB13 strain was carried out. Additionally, RT-qPCR measurements of central carbon metabolism, phospholipid degradation genes and the evaluation of growth and metabolite concentration of key mutants were determined. Results indicate that an early loss of the chromosomal region of ~10 kb is also present in the PB13 strain. In addition, two genes with punctual non-synonymous mutations in dhaM and glpT are present in both PB12 and PB13 strains. It is possible that after the occurrence of these genetic changes, a divergence event took place between the two PTS derivative strains. Derived from these changes we propose a novel metabolic strategy to contend with carbon stress that takes advantage of the degradation PtdEtn pathway in these strains. We also describe and compare the genetic basis of growth adaptation of PB12 and PB13 derivative E. coli variants, resulted from the same short term ALE process, that shows differential yields in the production of aromatic compounds, derived from specific genetic changes.
Results and discussion
Detection and description of the genetic changes in the PB12 and PB13 strains
Comparative genome analysis of the PB13 strain carried out by Winter Genomics Inc., shows that similar to the PB12 strain genome, several point mutations that are present in the evolved PB12 and PB13 strains were generated during the ALE process compared to the PB11 parental strain.
Mutations that occurred in the evolved PB13 strain during the ALE process
(A) JM101 vs PB13 Non-synonymous point mutations (WG)
Hypothetical/putative/unknown function genes
Predicted Wzy protein involved in ECA polysaccharide chain elongation
Predicted transporter of the major facilitator superfamily (MFS) of transporters
Metabolism and transport
One of several 6-phospho-β-glucosidases in E. coli
Phosphodeoxyribomutase; catabolic enzyme of the pyrimidine deoxyribonucleosides degradation pathway
Dihydroxyacetone kinase subunit M, homologous to certain PTS components
Major uptake transporter for glycerol-3-phosphate, belongs to the major Facilitator Superfamily (MFS)
IspG, catalyzes the sixth step in the methylerythritol phosphate pathway
Penicillin-insensitive d-alanyl-d-alanine (DD) endopeptidase
ATP-binding lipopolysacchride transport complex
Xanthosine MFS transporter
YdcT is an ATP-binding component of a predicted spermidine/putrescine ABC transporter
FliN, is one of three components of the flagellar motor’s “switch complex”
DNA Polymerase I (Pol I)
(B) JM101 vs PB13 Deletions (WG)
Member of a second PTS chain involved in nitrogen metabolism
RNA pyrophosphohydrolase that initiates mRNA degradation by hydrolysis of the 5′-triphosphate end
dGATC endonuclease in the MutHLS complex, the methyl-directed mismatch repair pathway
Putative transport protein
Putative NAD(P)-linked reductase that acts in starvation-associated mutations
Lysophospholipid transporter (LplT)
2-acylglycerophosphoethanolamine acyltransferase/acyl-ACP synthetase
Small RNA that is involved in regulating the protein composition of the outer membrane
Small RNA that is involved in regulating the protein composition of the outer membrane
DNA-binding transcription factor; represses transcription of the operons involved in transport and catabolism of D-galactose
Non-synonymous point mutations in the evolved strains
From the 13 non-synonymous point mutations detected in the PB13 strain, 2 are located on genes with putative or unknown function, 9 on genes involved in metabolism or transport, and the last 2 mapped in genes involved in motility and its regulation (Table 1). One interesting mutation originated in the polA gene, which codes for DNA polymerase I that is involved in several DNA repair pathways [15–17]. This mutation resulted in a valine instead of an alanine codon at position 695 of the protein inside of the Klenow fragment, which is involved in proofreading activity (Table 1). This mutation is located in a non-conserved region [18, 19]; however, this polymerase changes its conformational state according to its different roles, implicating several amino acids [20, 21]. In that sense, it is possible that this modification could have an effect on the polymerase proofreading activity.
Two relevant mutations are located in the dhaM and glpT genes and were detected in both evolved strains. The dhaM gene codes for the DhaM protein, which is a component of the dihydroxyacetone kinase (DhaK). This component consists of three PTS homologous domains (EI, Hpr and the A domain of the EII protein) . The mutation generates an opal stop codon (TGA) instead of a tryptophan codon (TGG) at position 346 of the protein (Table 1), which results in a truncated protein in the EI homologous domain.
In the PB13 strain, the appearance of a point mutation was noted in the msbA gene, which codes for the lipopolysaccharide transporter MsbA. This protein is involved in the translocation of PtdEtn to the periplasmic side of the inner membrane once this molecule has been recycled from 2-acyl-GPE by the Aas protein during the 2-acyl-GPE cycle in the cytoplasm  (Fig. 2). The mutation is located at position 123 of the protein in the transmembrane domain  and resulted in the substitution of a threonine by an alanine residue. Depletion or loss of function of MsbA results in the accumulation of lipopolysaccharide and phospholipids in the inner membrane of E. coli . The relevance of this particular mutation on the physiology of this strain is in principle of low impact in the turnover of PtdEtn because the genes encoding two proteins (Aas and LplT) involved in the 2-acyl-GPE cycle were lost during the ALE process. However, its occurrence, together with the aas-lplT operon deletion (discussed later) suggests that the adaptation in the PB12 and PB13 strains was directed towards the loss of function of the 2-acyl-GPE cycle.
According with their functions is unlikely that the remaining genes in PB13 with non-synonymous point mutations have an impact on fitness (Table 1).
Chromosomal deletion during the short term ALE process
As mentioned and reported for the PB12 strain , the deletion of a 10,328 bp chromosomal fragment including 12 contiguous genes (rppH, ygdT, mutH, ygdQ, ygdR, tas, lplT, aas, omrA, omrB, ptsP and galR) was detected in both PB12 and PB13 evolved strains. The deletion limits mapped within the ptsP and galR genes, resulting in a fusion of the remaining segments of these two genes (Fig. 1b). Earlier work describing the effect on several central metabolic gene transcripts and its impact on the glycolytic pathway allowing for better glucose uptake due to the absence of the rppH gene has been published . The detection of the same missing region in the PB13 chromosome establishes the importance of the adaptive nature of this DNA fragment deletion and proves that both strains were derived from the same ancestor.
Evolutionary divergence between the PB12 and PB13 strains
During the ALE experiment, strain PB11 was grown in a batch culture fermentor containing M9 minimal medium supplemented with 2 g/L of glucose (Fig. 1a). Under these conditions, a selection pressure was generated, favoring mutants that acquired the capacity to grow faster than the original PB11 parental strain. The batch culture was maintained until the stationary phase before starting a continuous culture, in which fresh medium was fed at progressively higher dilution rates (Fig. 1a) [2, 6, 7]. This procedure allowed for the isolation of the mutants generated during the batch culture according to their growth rates. Because two strains were isolated at different dilution rates during the continuous culture (0.4, 0.6 and 0.8 h−1), we tested for chromosomal integrity in six strains, including PB12 and PB13. All clone variants lacked the same chromosomal fragment (Fig. 1c), indicating that all strains were derived from the same ancestor. This strongly suggests that the chromosomal deletion is an important early adaptive event in the evolutionary process of this PB11 strain. Interestingly, this same deletion was detected in other evolved strains derived from a duplicate of the ALE experiment using the same PB11 strain (data not shown).
Once the derivatives acquired a growth advantage due to the loss of this specific chromosomal fragment, the cells were able to increase their μ and continued to accumulate mutations with an increased capacity due to the absence of the mutH gene that was also lost in the deletion. Although the precise emergence of dhaM and glpT point mutations is unknown, it is probable that these mutations were generated after the chromosomal deletion event because the cell has to overcome the absence of its principal phospholipid turnover mechanism, and these mutations could have a significant role to contend with this lost capacity, as discussed later (Figs. 2, 3). In addition, we tested for the presence of dhaM and glpT mutations in all the other intermediate strains and found both in all of them (data not shown). The shared mutations in the PB12 and PB13 strains establish that the divergence between PB12 and PB13 took place after the chromosomal deletion and the appearance of the dhaM and glpT point mutation events, likely at early stages of the exponential growth phase.
Contribution of the absent chromosomal fragment on growth rate
Importance of the mutated glpT and dhaM genes in the evolved strains
Specific growth rates of the JM101, PTS− strains and their derivatives with important inactivated genes
We also evaluated the glpQ gene, which is co-transcribed with glpT and is involved in the same phospholipid degradation pathway (Figs. 2, 3). Very similar results to those observed with the elimination of glpT occurred. Elimination of the glpQ gene did not affect the growth of the parental JM101 and PB11 strains, but a 40 and 45 % of growth reduction was observed in the PB12 and PB13 evolved strains, respectively (Table 2). These results strongly suggest the participation of this pathway in the growth recovery of the evolved strains. In agreement with this, both evolved strains lack the aas-lplT operon that is involved in membrane phospholipid recycling, which supports this last hypothesis.
To evaluate the contribution of the second shared punctual non-synonymous mutation on the physiology of the PTS− strains, individual ΔdhaM mutants were generated in all strains (Table 2). The results showed a decrease in the μ of all the PTS− mutant strains, especially in the evolved ones. The strain PB11 had a slight decrease in its μ, which is not considered significant. However, the PB12 and the PB13 strains had decreases of 16 and 40 %, respectively (Table 2). These results suggest that the DhaK could play an important role in the growth of the evolved strains, particularly in the PB13 strain. Interestingly, this protein (similar to the GlpT transporter) is involved in the glycerol pathway, which suggests that the phospholipid-glycerol-3-P degradation pathway contributes to the fitness increase in both evolved strains (Fig. 3). Nevertheless, it is unclear how this mutation provides an advantage because this mutation should decrease the phosphorylation capacities of the protein, which in turn would reduce the conversion of dihydroxyacetone (DHA) to dihydroxyacetone phosphate (DHAP) that serves as an intermediary in the glycolytic pathway. Another important point is that under glycolytic conditions, wild type cells are not able to produce DHA because this compound is generated from glycerol and not from glucose. It is possible that the consumption of G3P from the phospholipid metabolism in the evolved strains could be related to this behavior, but we are unable to completely explain the possible effect on cell physiology.
Advantage and contribution of the absence of the aas-lplT operon on the growth rate of the evolved strains
The phospholipids in the membrane of E. coli are phosphatidylethanolamine (PtdEtn) (75–80 %), phosphatidylglycerol (PtdGro) (15–20 %), and cardiolipin (CL) (~6 %) [33, 34]. PtdEtn can be metabolized by two pathways. The first route involves the recycling of this molecule by the 2-acyl-GPE cycle, in which all of its principal constituents are involved in a cycle and are not degraded further. The second metabolic pathway involves the complete degradation of the phospholipid molecule to G3P, fatty acids, and ethanolamine (Figs. 2, 3) and that could function in both evolved strains.
This second pathway in wild type cells requires the sequential activity of the apolipoprotein N-acyltransferase Lnt or the phospholipase PldA to generate a 2-acyl-GPE (Fig. 3a) and the inner membrane phospholipase PldB to generate a glycerophosphodiester and free fatty acid molecules from 2-acyl-GPE (Fig. 3b). Later in this pathway, the GDP protein coded by the glpQ gene hydrolyzes glycerophosphodiester to generate ethanolamine and G3P (Fig. 3c). The G3P molecule is in turn internalized from the periplasmic space to the cytoplasm through the GlpT transporter that is coded by the glpT gene . The free fatty acids formed during PldB action are converted to CoA derivatives by the activity of fatty acyl-CoA synthetase (FadD) . The G3P molecule is then redirected by the cell towards glycolysis or the de novo phospholipid biosynthesis pathway. The wild type cell could send fatty acids towards β-oxidation or to the phospholipid synthesis pathway (Fig. 3c). E. coli can also use ethanolamine as source of carbon and nitrogen ; however, a transporter has not been described.
It was previously reported that E. coli mutants with completely inactivated lplT and aas genes have a diminished PtdEtn turnover, resulting in an increase of periplasmic 2-acyl-GPE that is approximately 1/3 higher in the ΔlplT mutants compared to the wild type strain. The accumulation of 2-acyl-GPE is higher in the Δaas ΔpldB double mutants , which suggests a central role for the proteins with lysophospholipase A2 activity (PldB) on the alleviation of the 2-acyl-GPE accumulation by the secondary degradation route. It has also been reported that Δaas mutants have an altered membrane phospholipid composition and accumulate both 2-acyl-GPE and acyl-phosphatidylglycerol (acyl-PtdGro) . The acyl-PtdGro molecule is the result of the addition of a fatty acid from 2-acyl-GPE to the molecule of PtdGro by the lysophospholipase A2 PldB protein, establishing an additional fate for the 2-acyl-GPE that is modified to glycerol-3-phosphoethanolamine (a substrate for the GDP enzyme) (Figs. 2, 3).
With these elements, it is likely that 2-acyl-GPE and/or acyl-PtdGro derived from PtdEtn could be accumulated in the periplasmic space in the PB12 and PB13 evolved strains compared to the wild type parental JM101 due to the absence of their recycling machinery (Aas and LplT). Therefore, to understand the lipid metabolism in the PB12 and PB13 strains, lipids were analyzed by thin-layer chromatography (TLC). The results showed no significant differences between all the strains (data not shown) suggesting the use of an alternative pathway in the evolved strains, thus preventing the accumulation of lyso-phospholipids species, and an important metabolic plasticity in these strains.
RT-qPCR mRNA expression values of central carbon metabolism, PtdEtn degradation, β-oxidation and phospholipid biosynthesis genes in the evolved strains
RT-qPCR analysis of the main central carbon metabolism genes in the PB13 strain was performed. The expression levels of the same genes in the PB11 and PB12 strains were published in previous studies [2, 6, 7]. Interestingly, both evolved strains had similar expression values in most of the genes involved in the metabolic pathways tested (Additional file 3: Figure S1; Additional file 4: Table S3). This indicates that during the early stages of the evolutionary process (minimum 120 h), the main genetic changes responsible for the rearrangement of the expression profile were fixed; among these are the chromosomal fragment deletion and the two non-synonymous point mutations in the glpT and dhaM genes.
Increased global concentrations of mRNAs were detected due to the absence of rppH in PB11ΔrppH, PB12 and PB13 strains [2, 6, 7]. However, these higher mRNA concentrations were less intense in both evolved strains (Additional file 3: Figure S1; Additional file 4: Table S3), which suggests that one or more mutations are contributing to these phenomena. Despite the similarities, the higher transcription values of the gene coding for the galactose permease (galP) in the PB13 strain should be noted because its value increased 38-fold, while the PB12 increase was approximately 13-fold compare to the parental JM101 strain. Considering the higher growth rate in the PB13 strain (approximately 10 % compared to PB12), it is likely that the increased transcription of the galP gene provides an increased glucose uptake in this strain .
RT-qPCR mRNA expression values of PtdEtn degradation, β-oxidation and phospholipid biosynthesis genes in the evolved strains
1.03 ± 0.05
1.09 ± 0.26
1.47 ± 0.03
2.43 ± 0.46
0.97 ± 0.08
1.44 ± 0.08
2.25 ± 0.24
3.87 ± 0.71
1.2 ± 0.26
0.66 ± 0.1
0.95 ± 0.26
1.44 ± 0.39
1.89 ± 0.35
1.90 ± 0.10
2.59 ± 0.40
2.22 ± 0.05
63.46 ± 7.17
1.09 ± 0.26
1.47 ± 0.03
2.43 ± 0.46
1.71 ± 0.30
2.85 ± 0.08
4.79 ± 1.10
0.74 ± 0.05
1.57 ± 0.28
1.65 ± 0.13
30.50 ± 0.21
6.83 ± 1.07
7.52 ± 0.71
33.13 ± 2.89
6.20 ± 1.03
4.12 ± 0.69
1.2 ± 0.26
0.66 ± 0.1
0.95 ± 0.26
9.12 ± 1.43
56.93 ± 1.75
13.68 ± 1.56
2.14 ± 0.07
1.34 ± 0.02
2.12 ± 0.53
2.66 ± 0.40
3.00 ± 0.16
4.48 ± 0.22
Fatty acid biosynthesis
1.08 ± 0.20
0.64 ± 0.04
0.62 ± 0.09
0.77 ± 0.05
0.78 ± 0.05
0.67 ± 0.10
0.64 ± 0.02
0.42 ± 0.04
0.23 ± 0.05
2.25 ± 0.18
3.03 ± 0.16
4.21 ± 0.11
1.99 ± 0.22
1.26 ± 0.12
2.63 ± 0.33
1.25 ± 0.03
1.50 ± 0.05
1.95 ± 0.11
0.66 ± 0.06
1.54 ± 0.05
2.88 ± 0.04
0.69 ± 0.01
1.38 ± 0.11
1.74 ± 0.07
1.02 ± 0.19
1.30 ± 0.24
2.05 ± 0.21
2.04 ± 0.13
5.01 ± 0.20
3.00 ± 0.00
0.90 ± 0.06
1.72 ± 0.11
2.12 ± 0.28
0.59 ± 0.07
1.84 ± 0.01
1.91 ± 0.24
2.30 ± 0.13
2.12 ± 0.50
2.52 ± 0.31
0.52 ± 0.06
1.67 ± 0.12
2.35 ± 0.18
2.39 ± 0.15
2.34 ± 0.28
2.55 ± 0.08
Regarding the expression of genes involved in β-oxidation, high transcripts values were observed in all the strains, especially for the PB11 strain. This strain presents a permanent stress condition due to its inability for glucose uptake. As a result, carbon scavenging responses are triggered that allow the induction of several catabolic pathways [6–8]. As was presented in this contribution (Fig. 6), the PB11 strain accumulated significant levels of free fatty acids, even higher than those observed for PB12, suggesting that this scavenging strategy favors the free fatty acids reassimilation and further degradation in PB11 strain in order to increases the survival chances. It is important to mention that the inactivation of the aas-lplT operon in the PB11 strain allows a significant μ increase, which indicates that both, the 2-acyl-GPE cycle and the PtdEtn degradation pathways co-exist in this strain. In consequence, after the operon deletion, the 2-acyl-GPE cycle is no longer functional and all the carbon flow is directed through the degradation pathway to be reused as carbon molecules.
The high expression values observed for most of the genes coding for the free fatty acid degradation pathway in the evolved strains (Table 3), indicates as proposed, the reuse of free fatty acid molecules from the PtdEtn degradation pathway, which enhances their metabolism.
The expression values of the genes involved in the fatty acids biosynthetic pathway were also measured. The results show a diminished expression of most of these genes in the PB11 strain. On the other hand, both evolved strains present, high expression values (Table 3). These data, are in agreement with the proposal, since the absence of the 2-acyl-GPE cycle in the evolved strains, force the cells to replenish the fatty acids needed by restoring through this pathway the phospholipids required for a suitable performance. In this regard, it is possible to suggest that the utilization of both pathways in the PB11 results in a waste of carbon molecules, despite the further restoration of phospholipid molecules. Then, the loss of the 2-acyl-GPE cycle and the other mutations that allowed the degradation of PtdEtn in PB12 and the PB13 strains increased the fitness and the growth rate in these evolved derivatives.
Metabolic strategy in evolved strains lacking the ptsHIcrr, aas-lplT and galR genes
It is important to emphasize the pleiotropic effect of the absence of phospholipid turnover in the evolved and other strains. This absence is clearly advantageous under the PB11 PTS− scenario, where the strain is unable to use glucose efficiently, and a minimal increase in their ability to optimize carbon metabolism is manifested in its growth (PB11Δaas-lplT). In the evolved strains, this effect is not evident because both strains enhance glucose uptake through the effects of the rppH and galR deletions and are not limited for growth on glucose as the PB11 strain . However, the induction of phospholipid degradation in these strains is apparently required for an increased growth rate, as shown by the inactivation of the fadD gene. The decrease in growth rate in the PB11Δaas-lplT-fadD strain also supports this proposal (Table 2).
Supported on the previous results, we propose the follow strategy carried out by the PB11Δaas-lplT, PB12 and PB13 strains to overcome the absence of the aas-lplT operon. The activity of Lnt, but mainly of PldA and PldB are required to degrade PtdEtn and 2-acyl-GPE into fatty acids and glycerol-3-phosphoethanolamine. This latter compound is accumulated in the periplasmic space due to the lack of the LplT transporter (Fig. 7). In a second step, GDP would be responsible for G3P generation from the glycerophosphodiester molecule. With these reactions, the periplasmic fatty acids and G3P concentrations increase while further transportation by the FadD and GlpT proteins, respectively, is carried out (Fig. 7).
Once inside the cell, G3P is available to be redirected towards the glycolytic pathway or to “novo” phospholipid biosynthesis. On the other hand, the free fatty acids are also metabolized by the cell and are used to synthesize phospholipids or are degraded during β-oxidation (Fig. 7). This is possible in these strains because of the absence of catabolite repression due to the lack of the PTS system. Therefore, the evolved strains gain the capacity to recycle all the carbon molecules from phospholipid degradation to de novo phospholipid biosynthesis, glycolysis, and β-oxidation. These seem to favor the metabolism of evolved strains, allowing for a growth increase and, consequently, resulting in an increased fitness.
Essential genetic features for the optimization in the aromatic biosynthesis pathway in the PB12 strain
The ptsHIcrr deletion to increase the PEP availability is a widely employed strategy in metabolic engineering. The pleiotropic effect from that deletion has been analyzed from several approaches, in order to find a rational solution to overcome, essentially, the limited growth of the PTS− strains. The use of ALE as a tool in this field has allowed to solving this problem, however the appearance of undesirable genetic changes that could bring another unpredictable detrimental effects is an important issue. The evolved strains generated here and its further analysis allowed identifying the main mutations responsibles for uptake and metabolism of glucose. Interestingly, despite both evolved strains shared this mutations (rppH, galR, and aas-lplT deletions), further rational genetic modifications to divert the carbon flux towards the aromatic biosynthesis pathway result in a differential aromatic yields [40, 41]. The remarkable overproductive capacity of the PB12 over the PB13 strain was the result of some metabolic properties derived from specific point mutations in both strains.
The overproductive capacity in the PB12 strain comes from a particular uptake rate capacity (5.5 mmol glc h−1 gDW) and glucose catabolism that prevent an overflow through the glycolytic pathway, which in turn reduces the waste carbon in the acetic acid form . In addition, the strategies to redirect the carbon flow to the aromatic biosynthesis pathway through the overexpression of key genes diminish even more the acetic acid production [9, 10]. In the other hand, as result of the high levels of galP transcripts, the PB13 shows an increased uptake rate (8.5 mmol glc h−1 gDW) that generates the overflow responsible of the acetic acid production in similar concentrations compared to the wild type strain (56 % molar percentage) . This particularity in the PB13 strain represent a disadvantage for the aromatic compounds production, however, the genetic changes associated to the higher galP gene transcription is unknown.
Interestingly, some efforts in our laboratory to reproduce this metabolic behavior by reverse metabolic engineering approaches in other E. coli strains result in low aromatic compounds yields (data not shown), which indicate the importance of the genetic architecture. It seems the intrinsic genetic characteristics of the JM101 strain is desirable for this purpose, however the same fact imply that some other genetic backgrounds could be a better option for optimize the production of aromatic compounds.
The deletion of a chromosomal fragment containing important genes, such as rppH, mutH, galR and the aas-lplT operon, is proposed to be one of the early and most important adaptive events during the reported ALE process . The increase of mRNA levels of glycolytic genes because of the RppH pyrophosphohydrolase activity deficiency and the increased glucose uptake through GalP due to the lack of GalR [2, 6–8], in the PB12 and PB13 PTS− evolved strains allows to enhance their glycolytic metabolism. However, despite this growth advantage, the cells also lost their primary membrane phospholipid turnover pathway coded by the aas-lplT operon. The evolved strains also share two additional point mutations in the glpT and dhaM genes.
As the inactivations in the glpT and fadD genes suggests, the lack of the PtdEtn turnover metabolism capacity in the evolved strains was replaced with the utilization of the degradation pathway where phospholipids are transformed to their basic components: G3P and free fatty acids. The increase of more than 50 % in the μ of the derivative PB11Δaas-lplT strain supports the adaptive nature of this metabolic pathway blockage. In this situation, the evolved strains additionally improved carbon metabolism by forcing the cell to synthesize fatty acids de novo and taking advantage of the energy generated by β-oxidation and glycolysis during the degradation of fatty acids and G3P, respectively. Ultimately, the balance between de novo synthesis, β-oxidation, and glycolysis seems to be favorable in both PTS− PB12 and PB13 evolved strains, allowing for an increase in their growth rates on glucose as the only carbon source. These results indicate an important metabolic plasticity of these strains that lack PTS to utilize additional mutations (in glpT and dhaM genes) that occurred in the ALE to transform the deletion disadvantage of the aas-lplT operon lost in an advantage for growing faster on glucose by utilizing the capacity of recycling phospholipids as additional carbon and energy resources (Fig. 7). The RT-qPCR values of the genes involved in the capacity of recycling phospholipids that in general are overexpressed in the PB12 and PB13 evolved strains, are in agreement with this proposed recycling capacity that is enhanced and functional in the evolved strains and mainly in PB13 (Table 3).
The high increase in the growth rate observed with the deletion of rppH, aas, lplT and the entire chromosomal region in PB11 derivatives indicates that the contribution to the growth recovery by point mutations is minimal compared to the benefits provided by the chromosomal fragment deletion. However, there is a growth rate difference between both evolved strains caused by the higher galP gene expression in the PB13 strain (38-fold increase). It is possible that one of the point mutations present in the PB13 strain could have an indirect role in the increased galP transcription observed in this strain. In future experiments, we will individually inactivate these genes to see their effect on galP expression and growth rate.
The enormous stress caused by the elimination of PTS triggered a chromosomal rearrangement where the DNA mismatch repair system was affected due to the lack of mutH in the deleted DNA fragment. Consequently, the mutation rate increased. The appearance of a mutator population in bacterial cultures is a frequent event during limited growth conditions to improve its chances of survival by generating variation. In these PTS− evolved strains, this genetic novelty displaces alternative subpopulations because it creates in the resulting variants an improved capacity to metabolize glucose due to the absence of the rppH, aas, lplT and galR genes, ultimately increasing the fitness of the evolved strains. Interestingly, the absence of this specific chromosomal fragment could have given the cell the capacity to generate variation to explore several adaptive solutions that could overcome the lack of PTS.
Several ALE experiments have demonstrated the efficiency and how fast the evolution works in E. coli [2, 5, 42, 43]. This intrinsic feature has been exploited in biotechnology, resulting in a common practice during metabolic engineering strategies to improve particular features in the cell for enhance several bioprocess. In the PB12 and PB13, it was demonstrated that a particular qsGlc derived from specific genetic changes is the decisive factor to determine the efficiency in the aromatic compounds yields. Additionally, some results of preliminary reverse-metabolic engineering approaches to increase the glucose uptake in overproductive aromatic compounds PTS− strains, shows that the genetic architecture of the cell is essential. This issue it should be taking into account in order to enhance the productive capacities of the different metabolites in the industry.
The appearance of a mutator genotype in this kind of metabolic engineering strategies (ALE), allows the cell population to explore different genetic solutions. However, this genetic change in conjunction with the other deleted genes (galR, rppH, aas and lplT) resulted in the solution for an increased growth rate. At the moment after the deletion occurred, the cell culture could explore ways to diminish its mutation rate to decrease the probability of gaining deleterious mutations [44, 45], a fact that was not detected in both isolated evolved strains. Rather than decreasing, the PB13 strain increased its mutation rate. Among the genetic changes that could be involved in this increase is the polA mutation, which has a role in processing Okazaki DNA fragments and in filling gaps during the excision-repair processes. It is possible that this mutation could affect the performance of the DNA repair process. In the PB12 strain, where the mutation rate is moderate, the chance of generating deleterious mutations is still present. For that reason, it is desirable for stability and production purposes to consider the return of the wild type mutH gene, particularly in those PB12 derivative strains that are used for aromatic compound production. Nevertheless, cells where deleterious mutation could have occurred were not detected because of their inability to reproduce. In fact, increasing mutation rates in strains lacking mutH allows the cells in this culture to more quickly test several genetic changes (120 h in the reported study) and their effect on survival in different derivative cells.
In fact, an alternative for metabolite production purposes is that the presence of a mutator gene could allow mutations that enhance the production of the specific product. We cannot rule out this possibility in some of the PB12 derivatives that have increased capacities for producing aromatic compounds. For this reason, the genome of some of these strains will be resequenced.
Materials and methods
Bacterial strains, growth conditions, and recombinant DNA techniques
E. coli strains JM101 [F’ traD36 proAB + lacIq lacZΔM15/supE thi Δ(lac-proAB) rpoS(33am)], PB11 [JM101Δ(ptsH, ptsI, crr):: kan], PB12 (PB11ev1, PTS− Glc+) and PB13 (PB11ev2, PTS− Glc+) derivatives have previously been described [2, 6–8]. The derivatives of these strains utilized in this report (rppH, aas, lplT, glpT, dhaM, glpQ and fadD genes were knockout inactivated) were obtained by the Datsenko and Wanner method  using the oligonucleotides listed in Table S4 that is available in the Additional file 3: Figure S1. All gene disruptions were confirmed by PCR (data not shown). For inocula preparations, strains stored at −72 °C in glycerol were inoculated into Luria broth (LB) for overnight growth.
The cultures of the PB12 and PB13 strains that were utilized for DNA preparation for genome sequencing were obtained from the original culture that was kept frozen in glycerol [2, 7]. For μ determinations, cells were grown in LB and inoculated into M9 minimal medium with 2 g/L of glucose as the only carbon source. When the cultures were growing exponentially, they were inoculated into the same prewarmed (50 mL) medium, using 250 mL baffled flasks at 37 °C and stirred at 300 rpm with a starting optical density at 600 nm (O.D.600nm) = 0.1. O.D.600nm values were measured using a Klett/Summerson photocolorimeter, model 800-3. All specific μ values presented in Table 2 and Fig. 5a are the averages of at least three independent cultures, each one in duplicate. For RNA isolation and RT-qPCR analyses, duplicate cultures were grown in 1 L fermentors using M9 medium with 2 g/L of glucose as the sole carbon source. Cultures were grown at 37 °C, stirred at 600 rpm, and with an air flow rate of 1 vvm with a starting O.D.600nm = 0.1. For RT-qPCR determinations, cells from the different fermentations were collected in the log phase at O.D.600nm = 1 .
DNA extraction from parental and evolved PB13 strains for genomic analysis
Two overnight cultures of the E. coli strains JM101, PB11, and PB13 were grown from their frozen original stocks in liquid LB medium. DNA was extracted with a maxiprep phenol extraction and ethanol precipitation method [2, 6, 47] and purified with the Pure Link PCR purification kit (Invitrogen, USA). Quality and quantity of extracted DNA was verified as recommended by Roche NimbleGen (RN) and by UNAM Massive DNA Sequencing Unit. All strains were sequenced with Illumina Inc. Gallx high throughput technology. The JM101, PB11 and PB13 strains were resequenced with Ion Torrent PGM technology. The nucleotide sequence of the PB12 strain has been previously reported .
Library constructions and sequencing
DNA samples from the JM101, PB11 and PB13 strains were submitted to the Massive DNA Sequencing Unit of UNAM for its paired-ended (PE) library construction and genome sequencing. PE library was constructed following Illumina Inc. recommendations, which were previously described for the nucleotide sequence determination of strain PB12 . The Ion Xpress™ Plus Fragment Library Kit was used for the library construction, and genome sequencing by Ion Torrent PGM was performed following recommendations from Life Technologies.
Genome de novo assembly and variant identification
The protocol used for the “de novo” assembly and variant identification from the data obtained by high throughput sequencing with the Illumina GAIIx method was previously described . Ion Torrent reads were trimmed based on their quality using the FASTX-Toolkit (Hannon Lab. FASTX-Toolkit.). These reads were mapped against the JM101 assembly using bwa 0.6.2 . Variants were called using Samtools 0.1.18  and compared with variant calls from the PB13 assembly. Only variants reported by both methods (de novo assembly and Ion Torrent read mapping) were reported. For most of the analyses, the Instituto de Biotecnología, UNAM cluster resources were used.
DNA sequencing of putative mutations by Sanger methodology
DNA regions containing putative mutations detected in the PB12 strain were previously reported . For the PB13 strain, the point mutations detected by variant identification were confirmed. PCR amplifications were performed using the oligonucleotide primers listed in Additional file 5: Table S4 (shown in Additional file 3: Figure S1), purified by the Pure Link PCR purification kit, and sequenced by the Sanger methodology with the Taq FS Dye Terminator Cycle Sequencing Fluorescence-Based Sequencing in a Perkin Elmer/Applied Biosystems Model 3730. Sequence differences in 14 of the mutations are shown in PB13 in Table 1a and were confirmed by examination of the trace data (data not shown).
Estimation of the mutation rate in the different E. coli strains
Determination of mutation rates was based on mutation at a single locus, rpsL (rifampicin-resistant mutants). Mutation rates were determined by a modified Luria-Delbrück fluctuation test [30, 31]. Briefly, 100 µL of 30 overnight cultures were independently mixed and plated in LB agar plates containing the antibiotic selector (200 µg/mL of streptomycin). Samples of 100 µL from a 10−6 dilution from five overnight cultures were used to estimate the total viable count in drug-free LB plates. The MSS-LME and the LC methods  were used to estimate the number of mutants.
RNA extraction, cDNA synthesis, and RT-qPCR analysis
As previously described, total RNA was isolated and purified using a modified hot phenol method reported elsewhere [2, 6, 7], and a RevertAid™ H minus First Strand cDNA Synthesis kit was used to synthesize cDNA according to the manufacturer’s instructions (Fermentas, Burlington, Canada). For each reaction, approximately 5 μg of RNA and a mixture of 10 pmol/mL of specific DNA reverse primers (b) for each measured gene were used. The nucleotide sequences of these genes in PB12 have been previously published [2, 6, 7] RT-qPCR was performed with the ABI Prism 7300 Real-Time PCR System (Perkin Elmer/Applied Biosystems, Foster City, CA) using the MaximaR SYBR Green/ROX qPCR Master Mix (2x) kit (Fermentas LifeSciences), and reaction conditions were previously described . For each gene, all experiments were performed in triplicate from two different fermentations to obtain very similar values (differences <0.3 SD). A non-template control reaction mixture was included for each gene. Standard curves were constructed to evaluate PCR efficiency, and all the qPCR assays showed a high efficiency of amplification (90–100 %) because the genes had R2 values above 0.9976 with slopes between −3.4 and −3.7. All RT-qPCR experiments were compliant with the MIQE (Minimum Information for Publication of Quantitative Real-Time PCR Experiments) guidelines . The quantification technique used to analyze the data was the 2−ΔΔCq method described by Livak and Schmittgen . Data were normalized using the ihfB gene as an internal control.
In vivo labeling of E. coli with [14C]acetate and analysis of lipid extracts
The lipid composition of the different E. coli strains was determined by triplicate following labeling with [1-14C]acetate (Perkin Elmer). LB cultures (1.5 mL) were inoculated to an initial optical density at 600 nm (OD600) of 0.1 from precultures grown in the same medium. After the addition of 1 μCi/mL [1-14C]acetate to each culture, they were incubated for 24, 48 or 72 h. The cells were harvested by centrifugation and resuspended in 100 μL of water. The lipids were extracted according to the method of Bligh and Dyer . The chloroform phase was used for lipid analysis by one-dimensional thin-layer chromatography (TLC) using high-performance TLC silica gel 60 plates (Merck) and ethyl acetate-hexane-acetic acid (60:40:5 [vol/vol/vol]) as the mobile phase. Two-dimensional TLC was performed as described previously . Radioactivity was detected using a Storm 820 PhosphorImager (Amersham Biosciences). Image analysis and signal quantification were carried out using ImageQuant TL (Amersham Biosciences).
CA carried out all molecular, genetic, growth determinations and fermentor cultures of the strains. CA and NF performed the RT-qPCR analysis. FR-M participated in the genome assembly and variant identification of the strains. CA and DSC perform all lipid quantification experiments. CA, AE and FB carried out the data analysis. CA, SBC, OG, GG, AE and FB conceived the study and designed the experiments. CA and FB wrote the paper. All authors read and approved the final manuscript.
We thank Paul Gaytán, Jorge Yáñez, and Eugenio López for the synthesis of oligonucleotides and Sanger DNA sequencing. Ricardo Grande was responsible for the complete genome sequence performed at Unidad Universitaria de Secuenciación Masiva de ADN, UNAM. We also thank Ramón de Anda, Georgina Hernández, Mercedes Enzaldo, and Isabel López for technical support. This work was supported by CONACYT grant 240519.
The funders had no role in study design, data collection and analyses, decision to publish, or in the preparation of the manuscript.
The authors have declared that no competing interests exist.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Lee D-H, Palsson BØ. Adaptive evolution of Escherichia coli K-12 MG1655 during growth on a nonnative carbon source, l-1,2-propanediol. Appl Environ Microbiol. 2010;76:4158–68.View ArticleGoogle Scholar
- Aguilar C, Escalante A, Flores N, de Anda R, Riveros-McKay F, Gosset G, Morett E, Bolívar F. Genetic changes during a laboratory adaptive evolution process that allowed fast growth in glucose to an Escherichia coli strain lacking the major glucose transport system. BMC Genom. 2012;13:385.View ArticleGoogle Scholar
- Jang S-H, Kim J, Kim J, Hong S, Lee C. Genome sequence of cold-adapted Pseudomonas mandelii strain JR-1. J Bacteriol. 2012;194:3263.View ArticleGoogle Scholar
- Blázquez J, Couce A, Rodríguez-Beltrán J, Rodríguez-Rojas A. Antimicrobials as promoters of genetic variation. Curr Opin Microbiol. 2012;15:561–9.View ArticleGoogle Scholar
- Minty JJ, Lesnefsky AA, Lin F, Chen Y, Zaroff TA, Veloso AB, Xie B, McConnell CA, Ward RJ, Schwartz DR, et al. Evolution combined with genomic study elucidates genetic bases of isobutanol tolerance in Escherichia coli. Microb Cell Fact. 2011;10:18.View ArticleGoogle Scholar
- Flores N, Flores S, Escalante A, de Anda R, Leal L, Malpica R, Georgellis D, Gosset G, Bolívar F. Adaptation for fast growth on glucose by differential expression of central carbon metabolism and gal regulon genes in an Escherichia coli strain lacking the phosphoenolpyruvate:carbohydrate phosphotransferase system. Metab Eng. 2005;7:70–87.View ArticleGoogle Scholar
- Flores N, Leal L, Sigala JC, de Anda R, Escalante A, Martínez A, Ramírez OT, Gosset G, Bolivar F. Growth recovery on glucose under aerobic conditions of an Escherichia coli strain carrying a phosphoenolpyruvate:carbohydrate phosphotransferase system deletion by inactivating arcA and overexpressing the genes coding for glucokinase and galactose permease. J Mol Microbiol Biotechnol. 2007;13:105–16.View ArticleGoogle Scholar
- Flores S, Gosset G, Flores N, de Graaf AA, Bolı́var F. Analysis of carbon metabolism in Escherichia coli strains with an inactive phosphotransferase system by 13C labeling and NMR spectroscopy. Metab Eng. 2002;4:124–37.View ArticleGoogle Scholar
- Escalante A, Calderón R, Valdivia A, de Anda R, Hernández G, Ramírez OT, Gosset G, Bolívar F. Metabolic engineering for the production of shikimic acid in an evolved Escherichia coli strain lacking the phosphoenolpyruvate: carbohydrate phosphotransferase system. Microb Cell Factories. 2010;9:21.View ArticleGoogle Scholar
- Rodriguez A, Martínez JA, Báez-Viveros JL, Flores N, Hernández-Chávez G, Ramírez OT, Gosset G, Bolivar F. Constitutive expression of selected genes from the pentose phosphate and aromatic pathways increases the shikimic acid yield in high-glucose batch cultures of an Escherichia coli strain lacking PTS and pykF. Microb Cell Factories. 2013;12:86.View ArticleGoogle Scholar
- Rodriguez A, Martinez JA, Flores N, Escalante A, Gosset G, Bolivar F. Engineering Escherichia coli to overproduce aromatic amino acids and derived compounds. Microb Cell Factories. 2014;13:126.Google Scholar
- Sigala JC, Flores S, Flores N, Aguilar C, de Anda R, Gosset G, Bolivar F. Acetate metabolism in Escherichia coli strains lacking phosphoenolpyruvate: carbohydrate phosphotransferase system; evidence of carbon recycling strategies and futile cycles. J Mol Microbiol Biotechnol. 2009;16:224–35.View ArticleGoogle Scholar
- Flores N, de Anda R, Flores S, Escalante A, Hernández G, Martínez A, Ramírez OT, Gosset G, Bolívar F. Role of pyruvate oxidase in Escherichia coli strains lacking the phosphoenolpyruvate:carbohydrate phosphotransferase system. J Mol Microbiol Biotechnol. 2004;8:209–21.View ArticleGoogle Scholar
- Harvat EM, Zhang Y-M, Tran CV, Zhang Z, Frank MW, Rock CO, Saier MH. Lysophospholipid flipping across the Escherichia coli inner membrane catalyzed by a transporter (LplT) belonging to the major facilitator superfamily. J Biol Chem. 2005;280:12028–34.View ArticleGoogle Scholar
- Curti E, McDonald JP, Mead S, Woodgate R. DNA polymerase switching: effects on spontaneous mutagenesis in Escherichia coli. Mol Microbiol. 2009;71:315–31.View ArticleGoogle Scholar
- Makiela-Dzbenska K, Jaszczur M, Banach-Orlowska M, Jonczyk P, Schaaper RM, Fijalkowska IJ. Role of Escherichia coli DNA polymerase I in chromosomal DNA replication fidelity. Mol Microbiol. 2009;74:1114–27.View ArticleGoogle Scholar
- Hastings PJ, Hersh MN, Thornton PC, Fonville NC, Slack A, Frisch RL, Ray MP, Harris RS, Leal SM, Rosenberg SM. Competition of Escherichia coli DNA polymerases I, II and III with DNA Pol IV in stressed cells. PLoS ONE. 2010;5:e10862.View ArticleGoogle Scholar
- Dzidic S, Radman M. Genetic requirements for hyper-recombination by very short patch mismatch repair: involvement of Escherichia coli DNA polymerase I. Mol Gen Genet MGG. 1989;217:254–6.View ArticleGoogle Scholar
- Orren DK, Selby CP, Hearst JE, Sancar A. Post-incision steps of nucleotide excision repair in Escherichia coli. Disassembly of the UvrBC-DNA complex by helicase II and DNA polymerase I. J Biol Chem. 1992;267:780–8.Google Scholar
- Beese LS, Friedman JM, Steitz TA. Crystal structures of the Klenow fragment of DNA polymerase I complexed with deoxynucleoside triphosphate and pyrophosphate. Biochemistry (Mosc). 1993;32:14095–101.View ArticleGoogle Scholar
- Dzantiev L, Romano LJ. A conformational change in E. coli DNA polymerase I (Klenow fragment) is induced in the presence of a dNTP complementary to the template base in the active site. Biochemistry (Mosc). 2000;39:356–61.View ArticleGoogle Scholar
- Gutknecht R, Beutler R, Garcia-Alles LF, Baumann U, Erni B. The dihydroxyacetone kinase of Escherichia coli utilizes a phosphoprotein instead of ATP as phosphoryl donor. EMBO J. 2001;20:2480–6.View ArticleGoogle Scholar
- Larson TJ, Ehrmann M, Boos W. Periplasmic glycerophosphodiester phosphodiesterase of Escherichia coli, a new enzyme of the glp regulon. J Biol Chem. 1983;258:5428–32.Google Scholar
- Marger MD, Saier MH. A major superfamily of transmembrane facilitators that catalyse uniport, symport and antiport. Trends Biochem Sci. 1993;18:13–20.View ArticleGoogle Scholar
- Lemieux MJ, Song J, Kim MJ, Huang Y, Villa A, Auer M, Li X-D, Wang D-N. Three-dimensional crystallization of the Escherichia coli glycerol-3-phosphate transporter: a member of the major facilitator superfamily. Protein Sci Publ Protein Soc. 2003;12:2748–56.View ArticleGoogle Scholar
- Dawson RJP, Locher KP. Structure of a bacterial multidrug ABC transporter. Nature. 2006;443:180–5.View ArticleGoogle Scholar
- Doerrler WT, Raetz CRH. ATPase activity of the MsbA lipid flippase of Escherichia coli. J Biol Chem. 2002;277:36697–705.View ArticleGoogle Scholar
- Foster PL. Stress-induced mutagenesis in bacteria. Crit Rev Biochem Mol Biol. 2007;42:373–97.View ArticleGoogle Scholar
- Junop MS, Yang W, Funchain P, Clendenin W, Miller JH. In vitro and in vivo studies of MutS, MutL and MutH mutants: correlation of mismatch repair and DNA recombination. DNA Repair. 2003;2:387–405.View ArticleGoogle Scholar
- Ma WT, Sandri GV, Sarkar S. Analysis of the Luria-Delbrück distribution using discrete convolution powers. J Appl Probab. 1992;29:255–67.View ArticleGoogle Scholar
- Rosche WA, Foster PL. Determining mutation rates in bacterial populations. Methods San Diego Calif. 2000;20:4–17.View ArticleGoogle Scholar
- Hall BM, Ma C-X, Liang P, Singh KK. Fluctuation analysis CalculatOR: a web tool for the determination of mutation rate using Luria-Delbruck fluctuation analysis. Bioinforma Oxf Engl. 2009;25:1564–5.View ArticleGoogle Scholar
- Rilfors L, Lindblom G, Wieslander Å, Christiansson A. Lipid bilayer stability in biological membranes. In: Kates M, Manson LA (eds) Membrane fluidity. Springer, USA. 1984:205–245 [ Biomembranes , vol. 12].Google Scholar
- Morein S, Andersson A, Rilfors L, Lindblom G. Wild-type Escherichia coli cells regulate the membrane lipid composition in a “window” between gel and non-lamellar structures. J Biol Chem. 1996;271:6801–9.View ArticleGoogle Scholar
- Larson TJ, van Loo-Bhattacharya AT. Purification and characterization of glpQ-encoded glycerophosphodiester phosphodiesterase from Escherichia coli K-12. Arch Biochem Biophys. 1988;260:577–84.View ArticleGoogle Scholar
- Weimar JD, DiRusso CC, Delio R, Black PN. Functional role of fatty acyl-coenzyme A synthetase in the transmembrane movement and activation of exogenous long-chain fatty acids. Amino acid residues within the ATP/AMP signature motif of Escherichia coli FadD are required for enzyme activity and fatty acid transport. J Biol Chem. 2002;277:29369–76.View ArticleGoogle Scholar
- Garsin DA. Ethanolamine utilization in bacterial pathogens: roles and regulation. Nat Rev Microbiol. 2010;8:290–5.View ArticleGoogle Scholar
- Hsu L, Jackowski S, Rock CO. Isolation and characterization of Escherichia coli K-12 mutants lacking both 2-acyl-glycerophosphoethanolamine acyltransferase and acyl-acyl carrier protein synthetase activity. J Biol Chem. 1991;266:13783–8.Google Scholar
- Pech-Canul Á, Nogales J, Miranda-Molina A, Álvarez L, Geiger O, Soto MJ, López-Lara IM. FadD is required for utilization of endogenous fatty acids released from membrane lipids. J Bacteriol. 2011;193:6295–304.View ArticleGoogle Scholar
- Báez-Viveros JL, Osuna J, Hernández-Chávez G, Soberón X, Bolívar F, Gosset G. Metabolic engineering and protein directed evolution increase the yield of l-phenylalanine synthesized from glucose in Escherichia coli. Biotechnol Bioeng. 2004;87:516–24.View ArticleGoogle Scholar
- Báez-Viveros J, Flores N, Juárez K, Castillo-España P, Bolivar F, Gosset G. Metabolic transcription analysis of engineered Escherichia coli strains that overproduce l-phenylalanine. Microb Cell Factories. 2007;6:30.View ArticleGoogle Scholar
- Horinouchi T, Suzuki S, Hirasawa T, Ono N, Yomo T, Shimizu H, Furusawa C. Phenotypic convergence in bacterial adaptive evolution to ethanol stress. BMC Evol Biol. 2015;15:180.View ArticleGoogle Scholar
- Dragosits M, Mattanovich D. Adaptive laboratory evolution—principles and applications for biotechnology. Microb Cell Factories. 2013;12:64.View ArticleGoogle Scholar
- Turrientes M-C, Baquero F, Levin BR, Martínez J-L, Ripoll A, González-Alba J-M, Tobes R, Manrique M, Baquero M-R, Rodríguez-Domínguez M-J, Cantón R, Galán J-C. Normal mutation rate variants arise in a Mutator (Mut S) Escherichia coli population. PLoS ONE. 2013;8:e72963.View ArticleGoogle Scholar
- Galán J-C, Turrientes M-C, Baquero M-R, Rodríguez-Alcayna M, Martínez-Amado J, Martínez J-L, Baquero F. Mutation rate is reduced by increased dosage of mutL gene in Escherichia coli K-12. FEMS Microbiol Lett. 2007;275:263–9.View ArticleGoogle Scholar
- Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A. 2000;97:6640–5.View ArticleGoogle Scholar
- Ausubel FR. Short protocols in molecular biology. Fouth ed. New York: Wiley; 1999.Google Scholar
- Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinforma Oxf Engl. 2009;25:1754–60.View ArticleGoogle Scholar
- Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, Mueller R, Nolan T, Pfaffl MW, Shipley GL, Vandesompele J, Wittwer CT. The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin Chem. 2009;55:611–22.View ArticleGoogle Scholar
- Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods San Diego Calif. 2001;25:402–8.View ArticleGoogle Scholar
- Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification. Can J Biochem Physiol. 1959;37:911–7.View ArticleGoogle Scholar
- de Rudder KE, Thomas-Oates JE, Geiger O. Rhizobium meliloti mutants deficient in phospholipid N-methyltransferase still contain phosphatidylcholine. J Bacteriol. 1997;179:6921–8.Google Scholar