Extracellular Acidic pH Inhibits Acetate Consumption byDecreasing Gene Transcription of the Tricarboxylic Acid Cycleand the Glyoxylate Shunt

James S. Orr,a David G. Christensen,b Alan J. Wolfe,b Christopher V. Raoa

aDepartment of Chemical and Biomolecular Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois, USAbDepartment of Microbiology and Immunology, Stritch School of Medicine, Health Sciences Division, Loyola University Chicago, Maywood, Illinois, USA

ABSTRACT Escherichia coli produces acetate during aerobic growth on various car-bon sources. After consuming the carbon substrate, E. coli can further grow on theacetate. This phenomenon is known as the acetate switch, where cells transitionfrom producing acetate to consuming it. In this study, we investigated how pH gov-erns the acetate switch. When E. coli was grown on a glucose-supplemented me-dium initially buffered to pH 7, the cells produced and then consumed the acetate.However, when the initial pH was dropped to 6, the cells still produced acetate butwere only able to consume it when little (�10 mM) acetate was produced. Whensignificant acetate was produced in acidic medium, which occurs when the growthmedium contains magnesium, amino acids, and sugar, the cells were unable to con-sume the acetate. To determine the mechanism, we characterized a set of metabolicmutants and found that those defective in the tricarboxylic acid (TCA) cycle orglyoxylate shunt exhibited reduced rates of acetate consumption. We further foundthat the expression of the genes in these pathways was reduced during growth inacidic medium. The expression of the genes involved in the AckA-Pta pathway,which provides the principal route for both acetate production and consumption,was also inhibited in acidic medium but only after glucose was depleted, which cor-relates with the acetate consumption phase. On the basis of these results, we con-clude that growth in acidic environments inhibits the expression of the acetate ca-tabolism genes, which in turn prevents acetate consumption.

IMPORTANCE Many microorganisms produce fermentation products during aerobicgrowth on sugars. One of the best-known examples is the production of acetate byEscherichia coli during aerobic growth on sugars. In E. coli, acetate production is re-versible: once the cells consume the available sugar, they can consume the acetatepreviously produced during aerobic fermentation. We found that pH affects the re-versibility of acetate production. When the cells produce significant acetate duringgrowth in acidic environments, they are unable to consume it. Unconsumed acetatemay accumulate in the cell and inhibit the expression of pathways required for ace-tate catabolism. These findings demonstrate how acetate alters cell metabolism;they also may be useful for the design of aerobic fermentation processes.

KEYWORDS Escherichia coli, acetate, carbon metabolism

Many organisms, from bacteria to eukaryotes, exhibit overflow metabolism, whereby cellsgrowing aerobically on various carbon sources (e.g., glucose) produce a fermen-

tation product (1). While the primary fermentation products can vary among organisms(e.g., lactate for humans and acetate for Escherichia coli), the process itself occurs whenthe carbon flux through central metabolism exceeds the capacity of the tricarboxylicacid (TCA) cycle. To deal with this bottleneck, the cell diverts this excess flux to a

Citation Orr JS, Christensen DG, Wolfe AJ, RaoCV. 2019. Extracellular acidic pH inhibits acetateconsumption by decreasing gene transcriptionof the tricarboxylic acid cycle and theglyoxylate shunt. J Bacteriol 201:e00410-18.https://doi.org/10.1128/JB.00410-18.

Editor Conrad W. Mullineaux, Queen MaryUniversity of London

Copyright © 2018 American Society forMicrobiology. All Rights Reserved.

Address correspondence to Christopher V. Rao,cvrao@illinois.edu.

Received 10 July 2018Accepted 18 October 2018

Accepted manuscript posted online 22October 2018Published

RESEARCH ARTICLE

crossm

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 1Journal of Bacteriology

20 December 2018

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

fermentation product. This strategy enables cells to produce both energy and anabolicprecursors at high rates despite limitations in their metabolic capacity (for a review, seereference 2).

In many organisms, overflow metabolism can be reversed: once the cells consumethe available sugar, they will then consume the overflow metabolite to generate energyand provide anabolic precursors for additional growth. For example, during batchgrowth on glucose, E. coli consumes the acetate produced from overflow metabolism.Indeed, E. coli can grow on acetate as its sole carbon source (2). However, as discussedbelow, the ability of E. coli to consume acetate depends on the growth conditions,suggesting that this process involves more than simply cleaning up the leftovers.

As background, E. coli possesses three independent pathways for producing and/orconsuming acetate (Fig. 1). The primary pathway involves two enzymes: phosphotrans-acetylase (Pta) and acetate kinase (AckA). Pta catalyzes the reversible production ofacetyl-phosphate (acetyl-P) from acetyl coenzyme A (acetyl-CoA) and inorganic phos-phate. AckA catalyzes the reversible production of acetate and ATP from acetyl-P andADP. The second pathway involves pyruvate oxidase (PoxB), which catalyzes theirreversible decarboxylation of pyruvate into acetate and CO2. The third pathwayinvolves acetyl-CoA synthetase (Acs). Acs catalyzes the reversible production of acetyl-CoA, AMP, and pyrophosphate from acetate, ATP, and CoA. However, the reaction iseffectively irreversible because of the presence of intracellular pyrophosphatases. Onlythe Pta-AckA and PoxB pathways produce acetate, while only the Pta-AckA and Acspathways facilitate its consumption. The lower affinity Pta-AckA pathway provides thedominant mode for the consumption of high concentrations of acetate, whereas thehigher affinity Acs pathway provides the dominant mode for low concentrations ofacetate (3).

Despite extensive knowledge regarding the enzymes and pathways of acetatemetabolism, many questions remain regarding the regulation of these pathways. Inparticular, we still do not know all the factors involved in the differential regulation ofthese three pathways. Questions also remain concerning the consequences of acetatemetabolism at large. For example, acetyl-P, the intermediate of the Pta-AckA pathway,is the principal acetyl group donor for lysine acetylation, a global regulatory mechanismthat affects both the metabolism and physiology of the cell (3, 4). Acetyl-P is also knownto function as the phosphoryl donor to certain two-component response regulators (5).Understanding the regulation and its consequences is not simply an academic ques-tion, because it is often desirable to minimize acetate production in industrial appli-cations such as enzyme production without affecting cell viability (6, 7).

In this work, we investigated how the reduction in pH due to acetate production

FIG 1 Pathways of acetate metabolism. Acetate can be produced via the Pta-AckA pathway or PoxB andconsumed via the AckA-Pta pathway or Acs. Acetyl-CoA can then enter the TCA cycle or the glyoxylatebypass. Labeled genes are those used in this study.

Orr et al. Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 2

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

affects acetate consumption in E. coli. These studies were motivated by preliminaryexperiments showing that cells were unable to consume acetate when the growthmedium was weakly buffered to pH 6 or when the initial pH was not adjusted. Duringthese growth conditions, acetate accumulation acidified the medium to pH 5 or less. Toidentify the mechanism governing this behavior, we characterized acetate productionand consumption in a number of mutants lacking different metabolic enzymes. On thebasis of the behavior of these mutants and corroborating measurements of geneexpression, we found that the inability of the cells to consume acetate at low pH wasdue primarily to the reduced transcription of the genes that encode the tricarboxylicacid cycle and glyoxylate shunt.

RESULTSAcetate consumption is inhibited when the extracellular pH is acidic. When E.

coli was grown in tryptone broth buffered with an initial pH of 7 and containing22.2 mM (4 g/liter) glucose and 1 mM MgSO4 (TB7/gl�Mg), the cells consumed all ofthe sugar within the first 6 h (Fig. 2A). During this phase, acetate was rapidly produced,with a peak concentration at 6 h, the time when glucose was completely consumed.The acetate was then consumed, with most gone after 10 h of growth. At this point, cellgrowth ceased.

FIG 2 Low pH prevents acetate consumption from occurring. E. coli (MG1655) was grown aerobically for24 h in TB7 supplemented with glucose and magnesium (A), TB6 supplemented with glucose andmagnesium (B), and TB supplemented with glucose and magnesium (C).

Acidic pH Inhibits Acetate Consumption Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 3

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

When E. coli was grown in same medium with an initial pH of 6 (TB6/gl�Mg), thecells consumed the glucose at a somewhat lower rate than in TB7/gl�Mg (Fig. 2B).Acetate also was produced during this phase of growth, again rising rapidly until theglucose was completely consumed. However, the extracellular acetate concentrationdid not subsequently decrease, indicating that the cells did not consume it. Consistentwith the lack of acetate consumption, cell growth ceased at an optical density (OD) of�5 when the glucose was depleted from the growth medium. After cell growth ceased,the cells continued to produce acetate. The source of this additional acetate productionis unknown but is possibly a by-product of amino acid metabolism.

Similar growth experiments were also performed with the initial pH of 5 (see Fig.S1A in the supplemental material). Once again, the cells were unable to consumeacetate. In this case, however, the rate of glucose consumption was even lower, taking�24 h for the sugar to be consumed; cell growth ceased well before glucose depletion.Similar results were obtained when the growth medium was not buffered (TB/gl�Mg).The rate of glucose consumption was low, cell growth ceased long before glucosedepletion, and the cells did not consume acetate (Fig. 2C and S1B). These behaviorslikely resulted from the reduction in pH (�5) associated with acetate production (seeFig. S2).

Magnesium supplementation is necessary for robust glucose consumption in buff-ered tryptone broth containing glucose (8). In the absence of magnesium, the rate ofglucose consumption is much lower, with most occurring during stationary phase.Furthermore, when the medium lacks magnesium, the cells produce far less acetate. Inthe experiments described above, the medium was supplemented with magnesium. Todetermine if magnesium affects acetate consumption, the cells were grown in TB7/gland TB6/gl, which lack magnesium. In both cases, the cells consumed the acetate (seeFig. S3). Under these conditions, the production of acetate is less than that observed inmagnesium-supplemented conditions. As reported previously (8), these cultures pro-duced less acetate than those containing magnesium (TB7/gl, �10 versus �25 mM;TB6/gl, �8 versus �34 mM).

To determine whether these observations were unique to buffered tryptone broth,we also performed similar experiments in M9 minimal medium containing 22.2 mMglucose (see Fig. S4). At initial pH values of 6 and 7, very little acetate was produced(�2.5 mM), and in both conditions, the acetate was then consumed. However, whenthe M9 minimal medium containing 22 mM glucose was supplemented with CasaminoAcids (M9/gl�CA), the results were similar to those with buffered tryptone brothcontaining magnesium. Most importantly, cultures in M9/gl�CA produced significantlymore acetate (�17 mM) than cultures in M9/gl lacking Casamino Acids, suggesting theamino acids are necessary for high levels of acetate production. Consistent with acetateproduction being a by-product of overflow metabolism, the cultures with CasaminoAcids grew significantly faster than those without.

Acetate is also produced during growth on other sugars. To test whether the samebehavior is observed on sugars other than glucose, we grew E. coli in buffered tryptonebroth containing 1 mM MgSO4 using 11.7 mM (4 g/liter) lactose (see Fig. S5) or 26.6 mM(4 g/liter) arabinose (see Fig. S6) as the sugar. In both cases, the cells were able toconsume the acetate when the initial pH was 7 but not 6. We also tested whetheracetate was produced and then consumed when E. coli was grown in buffered tryptonebroth containing no sugar (Fig. 3). The cells produced little acetate (�2 mM). The cellswere also able to consume the acetate when the medium was initially buffered at pH7 or 6. Similar results were also observed when the growth medium containedmagnesium (data not shown). The ability to consume acetate in this case was likely dueto little acetate being produced, similar to what was observed with M9 medium lackingCasamino Acids and TB6/gl lacking magnesium.

Acetate metabolism is inhibited when the extracellular pH is acidic. To deter-mine whether the inhibition of acetate consumption was linked to sugar and/or aminoacid metabolism or to stationary-phase growth, we tested the ability of E. coli to

Orr et al. Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 4

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

consume acetate in tryptone broth containing 67.8 mM (4 g/liter) acetate (TB7/ac�Mgand TB6/ac�Mg) (Fig. 4). Consistent with the other growth experiments, the cells wereable to consume the acetate during growth in TB7/ac�Mg but not in TB6/ac�Mg.These results indicate that the inhibition of acetate utilization at low pH was due notto the available carbon sources (glucose and amino acids) themselves but rather to thepresence of substantial amounts of acetate that were produced from these carbonsources. Collectively, these results suggest that an extracellular environment combininglow pH and acetate inhibits acetate consumption.

Complete consumption requires both the Pta-AckA and Acs pathways. We nextexamined the pathways involved in acetate consumption by selectively eliminating thethree pathways involved in acetate metabolism. We first examined the behavior ofthese mutants during growth in TB7/gl�Mg (Fig. 5A, C, and E). The Δacs and ΔpoxBmutants produced similar amounts of acetate as the wild type (Fig. 5E). However, theΔacs mutant was unable to completely consume all of the acetate, with residual acetateconcentrations of approximately 3.5 mM. These results are not surprising, as the Acspathway is thought to be the principal route for acetate consumption at low acetateconcentrations (�7 mM) (2). The ΔpoxB mutant was able to completely consume theacetate; however, the rate of consumption was lower than the wild type. This slownesssuggests that PoxB exerts an unknown and indirect effect on acetate consumption, asthis pathway is irreversible (Fig. 1) and therefore is not directly involved in acetateconsumption. Both the ΔackA-pta and ΔackA-pta Δacs mutants produced acetate moreslowly than the wild type, with production peaking after 10 h of growth (Fig. 5E). Theseresults are consistent with the PoxB pathway being active after the transition tostationary phase (9). The ΔackA-pta mutant was able to slowly consume the acetate,whereas the ΔackA-pta Δacs mutant was not. The latter results are consistent with Acsbeing involved in the consumption of acetate produced from PoxB.

FIG 3 TB7 and TB6 without supplementation. Cell density and acetate accumulation in E. coli (MG1655)grown aerobically for 24 h in TB7 (solid lines) or TB6 (dashed lines) without glucose and magnesium.

FIG 4 TB7/ac�Mg and TB6/ac�Mg. Cell density and acetate accumulation in E. coli (MG1655) grownaerobically for 24 h in TB7 supplemented with acetate and magnesium (solid lines) or TB6 supplementedwith acetate and magnesium (dashed lines).

Acidic pH Inhibits Acetate Consumption Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 5

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

We also tested the ability of the ΔackA-pta Δacs mutant to grow on M9 mediumcontaining 16.9 mM (1 g/liter) acetate as the sole carbon source (see Fig. S7). Themutant was unable to grow, consistent with these two pathways being required foracetate consumption. When the mutant was complemented with a plasmid expressingackA-pta or acs, the cells were able to grow.

When all three pathways were deleted (ΔackA-pta Δacs ΔpoxB), little acetate wasproduced (Fig. 5E); it also was not consumed, presumably because all of the acetateconsumption pathways had been deleted. The pathway generating this small amountof acetate is unknown, though it may be a product of acetoacetate metabolism (10).

All three of the mutants deficient in ackA and pta grew more slowly than the wildtype (Fig. 5A) and consumed glucose at a lower rate in TB7/gl�Mg (Fig. 5C). However,even accounting for differences in growth, mutants lacking ackA and pta producedsignificantly less acetate (see Fig. S9). On the basis of these results, we conclude that thePta-AckA pathway is the principal route for acetate production and for the majority ofconsumption, while Acs is required for complete consumption of acetate duringgrowth in TB7/gl�Mg.

We next explored the behavior of these mutants during growth in TB6/gl�Mg (Fig.5B, D, and F) and TB/gl�Mg (see Fig. S8). With the exception of the ΔackA-pta Δacs

FIG 5 Growth of acetate metabolism mutants in different media. Cells were grown aerobically for 24 hin TB7 supplemented with glucose and magnesium (A, C, and E) or TB6 supplemented with glucose andmagnesium (B, D, and F). (The line for the ΔackA-pta mutant is superimposed over that for the ΔacsΔackA-pta mutant until 10 h in panel E.)

Orr et al. Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 6

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

ΔpoxB mutant, all of the mutants produced significant amounts of acetate (�25.4 mM),and the acetate continued to accumulate in the medium throughout the experiment.The same trends were also observed when the results were normalized to account fordifferences in growth (see Fig. S9 to S11). Acetate production was delayed in theΔackA-pta and ΔackA-pta Δacs mutants like it was in TB7/gl�Mg. Acetate productionwas also substantially reduced in the ΔpoxB mutant, again suggesting that PoxB exertsa small indirect effect on the production of acetate. No mutant consumed acetate. Thethree mutants deficient in ackA and pta grew and consumed glucose more slowly thanthe strains with an intact Pta-AckA pathway.

These results indicate that the Pta-AckA pathway provides the route for the majorityof the acetate consumption and that growth in an extracellular environment thatcombines low pH and acetate inhibits acetate consumption through both the Pta-AckAand Acs pathways.

Active transport is not required for acetate consumption. We next deleted thegenes that encode the high-affinity acetate/succinate:H� symporter SatP and theacetate/glycolate:cation symporter ActP, the only known active transport systems foracetate in E. coli (11, 12). The aim was to determine if deletion of one or both activetransport systems inhibited the ability of E. coli to consume acetate at pH 7 (see Fig.S12). When grown in TB7/gl�Mg, the ΔsatP and ΔactP single mutants and the ΔsatPΔactP double mutant excreted and consumed acetate no differently than their wild-type parent. These results suggest that acetate excretion and consumption under theseconditions are due to passive diffusion across the membrane; however, it is possiblethat acetate excretion and consumption involve some unknown transporter.

Disrupting the TCA cycle and glyoxylate bypass reduces the rate of acetateconsumption. Both the TCA cycle and glyoxylate bypass are involved in acetatemetabolism. Therefore, we tested whether deleting key genes in these two pathwaysinhibited the ability of E. coli to consume acetate during growth in TB7/gl�Mg (Fig. 6).We focused on sdhA, which encodes one of the catalytic subunits of succinate dehy-drogenase involved in the TCA cycle, and aceA and aceB, which encode isocitrate lyaseand malate synthase A, respectively, the enzymes of the glyoxylate bypass (Fig. 1).

The rate of acetate consumption was reduced in the ΔsdhA and ΔaceB mutants;however, both completely consumed the acetate by 24 h of growth (Fig. 6). The rate ofacetate consumption was further reduced in the ΔaceA mutant, which did not com-pletely consume the acetate after 24 h of growth. Acetate consumption was furtherinhibited in a ΔsdhA ΔaceA double mutant, indicating that both the TCA cycle and theglyoxylate bypass contribute to acetate consumption. The deletion of acs (ΔsdhA ΔaceAΔacs) did not further inhibit acetate consumption, suggesting that flux though thePta-AckA pathway is blocked by the inability of the cell to metabolize acetyl-CoA. Thereduced consumption of acetate in these mutants is somewhat similar to that observedduring growth in TB6/gl�Mg. As expected, the aceA mutant did not grow in M9medium when acetate was the sole carbon source; in contrast, the aceB mutantexhibited some growth (data not shown), likely because of its isozyme encoded by glcB(13). Collectively, these results suggest that the expression of enzymes in the TCA cycleand glyoxylate bypass may be inhibited during growth at low pH. We also performeda growth experiment in TB7/ac�Mg with some of the more interesting mutants shownpreviously: the ΔsatP ΔactP, ΔaceA ΔsdhA, and ΔackA-pta Δacs mutants (Fig. 7). Asexpected, neither the ΔaceA ΔsdhA mutant nor the ΔackA-pta Δacs mutant consumedany of the acetate after 24 h of growth. These two mutants also showed growth defects,with the ΔaceA ΔsdhA mutant growing more slowly than the ΔackA-pta Δacs mutant.The ΔsatP ΔactP mutant behaved similarly to the wild type. Once again, the behaviorwas similar to that observed during growth in TB6/ac�Mg.

Expression of multiple genes involved in acetate consumption is reduced at pH6. Since the disruption of the TCA cycle reduced acetate consumption, we testedwhether the expression of the TCA cycle genes was repressed when cells were grownat low pH. We used reverse transcription-quantitative PCR (qRT-PCR) to compare the

Acidic pH Inhibits Acetate Consumption Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 7

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

expression of key genes involved in the TCA cycle and glyoxylate bypass at two timepoints during growth in TB7/gl�Mg and TB6/gl�Mg (Fig. 8). We harvested a referencesample during exponential phase (3 h), when both cultures were behaving similarly,and again during stationary phase (10 h), when there was a difference in acetateconsumption (compare Fig. 2A and B). Consistent with a mechanism involving reducedexpression of the genes involved in acetate metabolism, we observed reduced expres-sion of sdhD, which encodes one of the subunits of succinate dehydrogenase, and aceBduring growth in TB6/gl�Mg versus TB7/gl�Mg after 3 and 10 h of growth. Presum-ably, aceA expression was also reduced, as it resides downstream of aceB in the sameoperon. We also observed reduced expression of the other TCA cycle genes, with fumBand perhaps fumA, which encode two of three fumarase isozymes, being the soleexceptions after 3 h of growth.

We also found that the expression of ackA was unchanged after 3 h of growth butreduced after 10 h of growth. As both ackA and pta reside in the same operon (14),these results suggest that the expression of the Pta-AckA pathway is unchanged during

FIG 6 Mutants of the TCA cycle and glyoxylate bypass. Cells were grown aerobically for 24 h in TB7supplemented with glucose and magnesium. (A) Cell density. (B) Glucose concentration. (C) Acetateaccumulation.

Orr et al. Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 8

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

the acetate production phase of growth (3 h) but reduced during the acetate con-sumption phase of growth (10 h) at pH 6, which is consistent with our other data.

Altogether, these expression data suggest that the cells are unable to consumeacetate at pH 6 because the expression levels of the pathways involved in acetatemetabolism are reduced.

ArcA is not the principal factor preventing acetate consumption at low pH. Onepossibility is that acidic pH affects the expression or phosphorylation of ArcA, which isknown to regulate many genes in the TCA cycle and glyoxylate bypass (15, 16).Therefore, we tested the growth of a ΔarcA mutant in TB6/gl�Mg (see Fig. S13). Therate of growth and glucose consumption in this mutant were similar to those of the

FIG. 7 Mutants of acetate metabolism, acetate transport, and TCA cycle/glyoxylate shunt grown inTB7/ac�Mg. Cells were grown aerobically for 24 h in TB7 supplemented with acetate and magnesium.(A) Cell density. (B) Acetate accumulation.

FIG. 8 Transcription levels for the TCA cycle and glyoxylate shunt are reduced in TB6 relative to that inTB7. E. coli (MG1655) was grown aerobically for 10 h in TB7 supplemented with glucose and magnesiumor TB6 supplemented with glucose and magnesium. After 3 h (black bars) and 10 h (gray bars), total RNAwas extracted and analyzed for the mRNA levels of the indicated genes, with 16S as a housekeepinggene. Relative expression was calculated as described in Materials and Methods. Values were analyzedby unpaired t tests. *, P � 0.05; **, P � 0.01; ***, P � 0.001; ****, P � 0.0001. Data are representative ofthree independent experiments and presented as the means � standard deviations (SDs).

Acidic pH Inhibits Acetate Consumption Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 9

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

wild type; however, it was still capable of consuming acetate. When we slightly reducedthe starting pH to 5.8, the ΔarcA mutant was unable to consume acetate. Likely, thebehavior at pH 6 was due to the production of less acetate by the mutant and,consequently, less acidification of its growth medium. This would suggest that ArcA isnot directly involved in inhibiting acetate consumption during growth in acidic media.However, it is also possible that ArcA blocks acetate consumption at pH 6 but anothermechanism further blocks it at lower pHs.

Succinate metabolism is not inhibited when the extracellular pH is acidic. Welast tested whether pH alters the consumption of succinate, whose extracellular con-centration can easily be measured. The goals of this experiment were to determinewhether the observed behavior is unique to acetate and whether the inhibition ofacetate consumption was solely due to pH. When we grew E. coli in buffered tryptonebroth containing magnesium and 33.9 mM succinate, we observed no difference in therate of utilization at either pH 6 or pH 7 (see Fig. S14). As succinate is metabolized bythe TCA cycle, these results suggest that the observed changes in metabolic geneexpression were not due to pH alone but also required acetate.

DISCUSSION

Many researchers have investigated different facets of acetate metabolism in E. coli(2). Much of this previous work was motivated by practical concerns, because acetateproduction can negatively impact cell viability and reduce product yields in fermenta-tion processes (6, 7). However, this previous work principally focused on the productionof acetate, with less attention devoted to acetate consumption. In the present study,we demonstrated that an acidic environment inhibits the ability of E. coli to consumeacetate during robust growth on sugars and amino acids. However, this inhibition wasonly observed when the environment also included significant (�10 mM) acetate,which occurs when the glucose-containing growth medium also contains magnesiumand amino acids. The magnesium and amino acids help support sufficiently highgrowth rates such that overflow metabolism, and thus acetate excretion, becomessignificant (8).

In the context of the present work, the previous study by Dittrich and coworkersmerits discussion (9). In that study, the authors investigated how pH affects acetateproduction in E. coli during growth in lysogeny broth containing glucose. Consistentwith our findings, the authors observed reduced expression of ackA at pH 6. Whereasthey observed reduced expression both after 3 h of growth and during stationaryphase, we observed reduced expression only during stationary phase, an observationconsistent with reduced acetate consumption but not reduced acetate productionduring growth at pH 6. While not discussed, their data also suggest that acetateconsumption is inhibited at pH 6; however, they did not consider the longer time pointsnecessary to verify a lack of consumption, as they focused their study on acetateproduction and not consumption. Finally, they also found that the PoxB pathway isactive after entry into stationary phase, which is consistent with our findings.

Why then is acetate consumption inhibited by low pH? Many bacteria produce andconsume organic acids. In these organisms, the pH of the growth medium is known toaffect their metabolism. While the mechanisms governing this phenomenon are gen-erally unknown, it is thought in part to reflect the energetics of acid transport (17). Ifonly the protonated acid is able to cross the plasma membrane, then the energetics willbe determined solely by the difference in the intra- and extracellular concentrations ofthe protonated acid, regardless of whether the transport mechanism is passive oractive. If there is also no pH difference across the cell membrane, then the total acidconcentrations will be the same unless an energy-dependent transporter is employedto sustain a concentration difference. However, if the extracellular pH is lower than theintracellular pH, then more of the extracellular acid will be in the protonated form. Thismeans that even if the total acid concentration is the same inside and outside the cell,there will be a concentration difference in protonated acid across the membrane,favoring acid import. In other words, it is energetically more expensive for the cell to

Orr et al. Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 10

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

excrete acid when the extracellular pH is acidic than when it is neutral. On the basis ofthese simple arguments, we would expect acid excretion when the extracellular pH isneutral and acid import when the extracellular pH is acidic.

Active transport mechanisms can overcome concentration differences by supplyingthe requisite energy for acid transport, even when the extracellular pH is acidic.However, acetic acid transport in E. coli is thought to involve a passive mechanism (18).While E. coli does possess two transporters for acetate uptake, SatP (acetate:H�

symporter) and ActP (acetate:Na� symporter), they are thought to be involved inuptake at low concentrations, far below those observed in our experiments (12).Indeed, deleting the genes encoding these two transporters did not affect the ability ofE. coli to consume acetate when the initial pH was 7 (see Fig. S12 in the supplementalmaterial). This suggests that acetate in its protonated form (acetic acid) is able topassively cross the membrane. It also suggests that acetate uptake is favored when theextracellular pH is acidic, contrary to what we observed.

So how then is acetate consumption stopped at low pH? The simplest explanation,which is consistent with our data, is that acetate accumulates in the cell during growthin acidic environments and inhibits the expression of many enzymes involved inacetate catabolism (Fig. 8). As a consequence, the cells are unable to consume theacetate. From simple equilibrium calculations (assuming an intracellular pH of 7),acetate is predicted to accumulate in the cell at pH 6 such that the intracellularconcentration of total acid is 9.5-fold greater than the extracellular concentration.Likely, this high concentration of intracellular acetate is what inhibits the expression ofthe genes involved in acetate catabolism. In support of this model, intracellular acetateis known to inhibit the basal transcription of many genes (19), which potentiallyexplains the observed changes in gene expression (Fig. 8). This model also explains whycells were able to consume acetate at pH 6 when grown in the absence of amino acids(Fig. S4), because the amount of acetate they produced was insufficient for acetate toaccumulate in the cells at sufficiently high concentrations to inhibit the expression ofthe acetate catabolic genes. A similar process likely occurs during growth at pH 7,where intra- and extracellular concentrations of total acid are expected to be moresimilar. Even though the cells produce significant concentrations of acetate whengrown in the presence of amino acids, the lack of a large pH gradient across the cellmembrane means that the acetate will not accumulate within the cells. In fact, theintracellular concentration at pH 7 is predicted to be 10-fold less than that at pH 6.

In conclusion, we found that E. coli is unable to consume acetate during growth inacidic environments. The behavior was observed during growth on peptide-basedmedia containing glucose, because the cells produce acetate as an overflow metabolitebut cannot consume it due to the repression of many genes involved in acetatecatabolism.

MATERIALS AND METHODSGrowth conditions, media, strains, and plasmids. All growth experiments were performed with

25 ml of liquid media in 250-ml flasks without baffles. The flasks were aerated by shaking at 225 rpm. Thefollowing media were used for cell growth: LB (10 g/liter tryptone, 10 g/liter NaCl, 5 g/liter yeast extract),TB (10 g/liter tryptone and 5 g/liter NaCl), TB7 (10 g/liter tryptone and buffered at pH 7.0 with 61.5 mMK2HPO4 and 38.2 mM KH2PO4), TB6 (10 g/liter tryptone and buffered at pH 6.0 with 13.2 mM K2HPO4 and86.8 mM KH2PO4), TB5 (10 g/liter tryptone, 100 mM KH2PO4, and pH adjusted to 5.0 with HCl), and M9minimal medium (6.78 g/liter Na2HPO4, 3 g/liter KH2PO4, 1 g/liter NH4Cl, 0.5 g/liter NaCl, 0.1 mM CaCl2,1 mM MgSO4, 2.45 �M ferric citrate, 0.03 mM thiamine, and pH adjusted to 6.0 or 7.0 with HCl). Additionalcarbon sources were added to the growth media at a concentration of 4 g/liter, unless noted otherwise.In media containing magnesium, MgSO4 was added to a final concentration of 1 mM. When appropriate,Casamino Acids were added to a final concentration of 1%.

All strains used in this study are listed in Table 1. Single gene deletions were constructed either byP1 transduction (20) to move deletions of interest from the corresponding strain in the Keio collection(21) or by the method of lambda Red recombination (22). The primers used for lambda Red-mediateddeletions are listed in Table S1 in the supplemental material. Double or multiple mutant strains weremade by transferring additional mutations into a desired mutant background using P1 transduction.Mutants were selected for by growth on LB plates supplemented with the appropriate antibiotic. PCRwas performed after all transduction or lambda Red steps and the results were compared to those fromthe wild type to verify the mutations.

Acidic pH Inhibits Acetate Consumption Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 11

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

The complementation plasmids pUC19::acs and pUC19::ackApta were constructed by separatelycloning the DNA fragment containing the ackA-pta and acs operons (with 300 bp of flanking sequence)into pUC19 using Gibson assembly (23). The DNA sequences for the ackA-pta and acs operons wereobtained from ecocyc.org (25). The primers used for plasmid construction are listed in Table S1 in thesupplemental material.

Analytical methods. Cell density was measured by the optical density at 600 nm. Glucose,arabinose, lactose, acetate, and succinate concentrations were measured using a Shimadzu high-performance liquid chromatography system equipped with an RID-10A refractive index detector, anAminex HPX-87H carbohydrate analysis column (Bio-Rad Laboratories, Hercules, CA), and a cation HMicro-Guard cartridge (Bio-Rad Laboratories). The column and guard cartridge were kept at 65°C,and 5 mM H2SO4 was used as the mobile phase at a constant flow rate of 0.6 ml/min. Prior to theanalysis, culture samples were first pelleted and then the supernatant was passed through 0.22-�mpolyethersulfone syringe filter. The peaks were identified and quantified by retention time compar-ison to authentic standards.

qRT-PCR. For the quantification of gltA, sucA, sucC, sdhD, acnB, acnA, icd, fumA, fumC, fumB, mdh,aceB, and ackA mRNA, the cells were aerated at 37°C in TB7 or TB6 supplemented with 0.4% glucose and1 mM MgSO4 and harvested after 3 h of growth (optical density at 600 nm [OD600] of 1.1 to 1.3) and instationary phase after 10 h. A volume of culture equivalent to 0.5 ml of culture at an OD600 of 1 wasadded to 1 ml RNAprotect bacterial reagent (76506; Qiagen) and then vortexed. After centrifugation, thecell pellet was frozen at �80°C and stored overnight. RNA was isolated using the MasterPure RNApurification kit (MCR85102; Epicentre) according to the manufacturer’s protocol. RNA concentration/purity was determined using a NanoDrop 2000 spectrophotometer (Thermo Scientific). Next, 7.5 �g or5 �g RNA was used to synthesize cDNA with the iScript cDNA synthesis kit (Bio-Rad 170 – 8890),according to the manufacturer’s protocol. Reverse transcription-quantitative PCR (qRT�PCR) was per-formed using iTaq universal SYBR green supermix (172-5120; Bio-Rad), according to manufacturer’sprotocol, on a CFX96 real-time system (Bio-Rad) with a C1000 thermal cycler (Bio-Rad) and the followingconditions: 95°C for 5 min, then 37 cycles of 95°C for 15 s, 60°C for 20 s, and 72°C for 30 s. The list ofprimers used can be found in Table S2. The quantitation of 16S rRNA was used to normalize the data.Standard curves of each primer set were generated with gDNA from E. coli strain B (D4889; Sigma), and theamplification efficiency (E) was determined (see Table S3). Relative expression was calculated as describedpreviously (24). Briefly, the ΔCT values [CT(TB7) � CT(TB6)] for each gene of interest (GOI) and the 16S house-keeping gene (HKG gene) were determined. Then, the gene expression ratio was determined with acorrection for amplification efficiency via the following equation: gene expression ratio � EGOI

ΔCT(GOI)/EHKG

ΔCT(HKG).

TABLE 1 Strains used in this study

Strain Genotype or description Source or reference

MG1655 F� �� ilvG rfb-50 rph-1 CGSC no. 7,740BW25113 F� Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) �� rph-1 Δ(rhaD-rhaB)568 hsdR514 CGSC no. 7,636DH5� F� Δ(argF-lac)169 �80dlacZ58(M15) ΔphoA8 glnX44(AS) �� deoR481 rfbC1?

gyrA96(NalR) recA1 endA1 thiE1 hsdR17CGSC no. 12,384

MG1655 pKD 46 Lambda Red recombinase 23DH5� pKD 3 Source of FRT-cat-FRT 23DH5� pKD 4 Source of FRT-kan-FRT 23JO71 MG1655 Δacs::FRT This studyJO73 MG1655 ΔackA-pta::FRT This studyJO74 MG1655 Δacs::FRT ΔackA-pta::kan This studyJO75 MG1655 Δacs::FRT ΔackA-pta::FRT This studyJO77 MG1655 ΔpoxB::cat This studyJO78 MG1655 Δacs::FRT ΔackA-pta::FRT ΔpoxB::cat This studyJW4028-1 BW25113 ΔactP::kan 22JO114 MG1655 ΔactP::kan This studyJO115 MG1655 ΔactP::kan ΔsatP::cat This studyJO116 MG1655 ΔsatP::cat This studyJW3975-3 BW25113 ΔaceA::kan 22JO147 MG1655 ΔaceA::FRT This studyJW3974-1 BW25113 ΔaceB::kan 22JO133 MG1655 ΔaceB::kan This studyJW0713-1 BW25113 ΔsdhA::kan 22JO144 MG1655 ΔsdhA::kan This studyJO164 MG1655 ΔaceA::FRT ΔsdhA::FRT This studyJO169 MG1655 ΔaceA::FRT ΔsdhA::FRT Δacs::kan This studyJW4364 BW25113 ΔarcA::kan 22JO193 MG1655 ΔarcA::kan This studyJO344 JO74 pUC19::ackApta This studyJO345 JO74 pUC19::acs This study

Orr et al. Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 12

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SUPPLEMENTAL MATERIAL

Supplemental material for this article may be found at https://doi.org/10.1128/JB.00410-18.

SUPPLEMENTAL FILE 1, PDF file, 5.8 MB.

ACKNOWLEDGMENTS

This material is based upon work supported by the U.S. Department of Energy,Office of Science, Office of Biological and Environmental Research, under award num-ber DE-SC0012443.

REFERENCES1. Vazquez A. 2017. Overflow metabolism: from yeast to marathon runners.

Academic Press, Cambridge, MA.2. Wolfe AJ. 2005. The acetate switch. Microbiol Mol Biol Rev 69:12–50.

https://doi.org/10.1128/MMBR.69.1.12-50.2005.3. Kuhn ML, Zemaitaitis B, Hu LI, Sahu A, Sorensen D, Minasov G, Lima BP,

Scholle M, Mrksich M, Anderson WF, Gibson BW, Schilling B, Wolfe AJ.2014. Structural, kinetic and proteomic characterization of acetylphosphate-dependent bacterial protein acetylation. PLoS One 9:e94816.https://doi.org/10.1371/journal.pone.0094816.

4. Weinert BT, Iesmantavicius V, Wagner SA, Scholz C, Gummesson B, BeliP, Nystrom T, Choudhary C. 2013. Acetyl-phosphate is a critical deter-minant of lysine acetylation in E. coli. Mol Cell 51:265–272. https://doi.org/10.1016/j.molcel.2013.06.003.

5. Wolfe AJ. 2010. Physiologically relevant small phosphodonors link me-tabolism to signal transduction. Curr Opin Microbiol 13:204 –209. https://doi.org/10.1016/j.mib.2010.01.002.

6. De Mey M, De Maeseneire S, Soetaert W, Vandamme E. 2007. Minimizingacetate formation in E. coli fermentations. J Ind Microbiol Biotechnol34:689 –700. https://doi.org/10.1007/s10295-007-0244-2.

7. Eiteman MA, Altman E. 2006. Overcoming acetate in Escherichia colirecombinant protein fermentations. Trends Biotechnol 24:530 –536.https://doi.org/10.1016/j.tibtech.2006.09.001.

8. Christensen DG, Orr JS, Rao CV, Wolfe AJ. 2017. Increasing growth yieldand decreasing acetylation in Escherichia coli by optimizing the carbon-to-magnesium ratio in peptide-based media. Appl Environ Microbiol83:e03034-16. https://doi.org/10.1128/AEM.03034-16.

9. Dittrich CR, Bennett GN, San KY. 2005. Characterization of the acetate-producing pathways in Escherichia coli. Biotechnol Prog 21:1062–1067.https://doi.org/10.1021/bp050073s.

10. Sramek SJ, Frerman FE. 1975. Purification and properties of Escherichiacoli coenzyme A-transferase. Arch Biochem Biophys 171:14 –26. https://doi.org/10.1016/0003-9861(75)90002-8.

11. Gimenez R, Nunez MF, Badia J, Aguilar J, Baldoma L. 2003. The gene yjcG,cotranscribed with the gene acs, encodes an acetate permease in Esch-erichia coli. J Bacteriol 185:6448 – 6455. https://doi.org/10.1128/JB.185.21.6448-6455.2003.

12. Sa-Pessoa J, Paiva S, Ribas D, Silva IJ, Viegas SC, Arraiano CM, Casal M.2013. SATP (YaaH), a succinate-acetate transporter protein in Escherichiacoli. Biochem J 454:585–595. https://doi.org/10.1042/BJ20130412.

13. Pellicer MT, Fernandez C, Badia J, Aguilar J, Lin EC, Baldom L. 1999.Cross-induction of glc and ace operons of Escherichia coli attributable topathway intersection. Characterization of the glc promoter. J Biol Chem274:1745–1752. https://doi.org/10.1074/jbc.274.3.1745.

14. Kakuda H, Hosono K, Shiroishi K, Ichihara S. 1994. Identification andcharacterization of the ackA (acetate kinase A)-pta (phosphotransacety-lase) operon and complementation analysis of acetate utilization by anackA-pta deletion mutant of Escherichia coli. J Biochem 116:916 –922.https://doi.org/10.1093/oxfordjournals.jbchem.a124616.

15. Iuchi S, Lin EC. 1988. arcA (dye), a global regulatory gene in Escherichiacoli mediating repression of enzymes in aerobic pathways. Proc NatlAcad Sci U S A 85:1888 –1892. https://doi.org/10.1073/pnas.85.6.1888.

16. Lynch AS, Lin EC. 1996. Transcriptional control mediated by the ArcAtwo-component response regulator protein of Escherichia coli: charac-terization of DNA binding at target promoters. J Bacteriol 178:6238 – 6249. https://doi.org/10.1128/jb.178.21.6238-6249.1996.

17. van Maris AJ, Konings WN, van Dijken JP, Pronk JT. 2004. Microbialexport of lactic and 3-hydroxypropanoic acid: implications for industrialfermentation processes. Metab Eng 6:245–255. https://doi.org/10.1016/j.ymben.2004.05.001.

18. Axe DD, Bailey JE. 1995. Transport of lactate and acetate through theenergized cytoplasmic membrane of Escherichia coli. Biotechnol Bioeng47:8 –19. https://doi.org/10.1002/bit.260470103.

19. Rosenthal AZ, Kim Y, Gralla JD. 2008. Regulation of transcription byacetate in Escherichia coli: in vivo and in vitro comparisons. Mol Microbiol68:907–917. https://doi.org/10.1111/j.1365-2958.2008.06186.x.

20. Thomason LC, Costantino N, Court DL. 2007. E. coli genome manipula-tion by P1 transduction. Curr Protoc Mol Biol Chapter1:Unit 1.17. https://doi.org/10.1002/0471142727.mb0117s79.

21. Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA,Tomita M, Wanner BL, Mori H. 2006. Construction of Escherichia coli K-12in-frame, single-gene knockout mutants: the Keio collection. Mol SystBiol 2:2006.0008. https://doi.org/10.1038/msb4100050.

22. Datsenko KA, Wanner BL. 2000. One-step inactivation of chromosomalgenes in Escherichia coli K-12 using PCR products. Proc Natl Acad SciU S A 97:6640 – 6645. https://doi.org/10.1073/pnas.120163297.

23. Gibson DG, Young L, Chuang RY, Venter JC, Hutchison CA, III, Smith HO.2009. Enzymatic assembly of DNA molecules up to several hundredkilobases. Nat Methods 6:343–345. https://doi.org/10.1038/nmeth.1318.

24. Pfaffl MW. 2001. A new mathematical model for relative quantification inreal-time RT-PCR. Nucleic Acids Res 29:e45. https://doi.org/10.1093/nar/29.9.e45.

25. Keseler IM, Collado-Vides J, Gama-Castro S, Ingraham J, Paley S, PaulsenIT, Peralta-Gil M, Karp PD. 2005. EcoCyc: a comprehensive databaseresource for Escherichia coli. Nucleic Acids Res 33:D334 –D337. https://doi.org/10.1093/nar/gki108.

Acidic pH Inhibits Acetate Consumption Journal of Bacteriology

January 2019 Volume 201 Issue 2 e00410-18 jb.asm.org 13

on Septem

ber 5, 2020 by guesthttp://jb.asm

.org/D

ownloaded from