Biochimica et Biophysica Acta 1833 (2013) 2879–2889

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbamcr

A role for EIIANtr in controlling fluxes in the central metabolism ofE. coli K12

Susan Jahn a, Bart R. Haverkorn van Rijsewijk b, Uwe Sauer b, Katja Bettenbrock a,⁎a Max Planck Institute for Dynamics of Complex Technical Systems, Sandtorstrasse 1, 39104 Magdeburg, Germanyb Institute for Molecular Systems Biology, ETH Zürich, Wolfgang-Pauli Str. 16, 8093 Zurich, Switzerland

⁎ Corresponding author. Tel.: +49 3916110249; fax: +E-mail address: bettenbrock@mpi-magdeburg.mpg.de

0167-4889/$ – see front matter © 2013 Elsevier B.V. All rhttp://dx.doi.org/10.1016/j.bbamcr.2013.07.011

a b s t r a c t

a r t i c l e i n f o

Article history:Received 21 May 2013Received in revised form 4 July 2013Accepted 15 July 2013Available online 23 July 2013

Keywords:Phosphotransferase systemMetabolic controlOverflow metabolismNitrogen PTS

To investigate a possible role of the nitrogen-PTS (PTSNtr) in controlling carbon metabolism, we determined thegrowth of Escherichia coli LJ110 and of isogenic derivatives, mutated in components of the PTSNtr, on differentcarbon sources. The PTSNtr is a set of proteins homologous to the PEP-dependent phosphotransferase system(C-PTS) that transfers a phosphate group from PEP over EINtr (encoded by ptsP) and NPr (encoded by ptsO) toEIIANtr (encoded by ptsN). Strains deleted in ptsN were characterized by a high acetate production coupled toslow growth on glycolytic substrates. The ΔptsP and the ΔptsO strain showed the same behavior as the parentstrain. As the phosphorylation level of EIIANtr in these mutants differed significantly from that of the parentstrain, phosphorylation of EIIANtr obviously is not important for its function. During growth in minimal mediumwith defined carbon sources, EIIANtr was always completely phosphorylated in LJ110. Significant amounts ofdephosphorylated EIIANtr were only visible in strains lacking EINtr or NPr. mRNA expression studies on glucoserevealed a downregulation of genes encoding TCA cycle enzymeswhenEIIANtrwas absent. 13C-flux analyses con-firmed higher fluxes towards acetate and lower fluxes in the TCA cycle in the ptsN mutants but additionallyhinted to a slightly but significantly increased flux through the pyruvate dehydrogenase complex (PDH). Duringgrowth on succinate the ΔptsN strain accumulated mutations in rpoS, while no rpoSmutants were observed forthe ΔptsN-O strain. This hints to an additional function of NPr during growth with succinate.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

The PEP dependent sugar phosphotransferase system (C-PTS)mediates the uptake of about 15 carbohydrates in Escherichia coli,the so-called PTS-substrates [1,2]. In the PTS uptake reactions, thephosphate from PEP is transferred via Enzyme I (EI, encoded by ptsI)to HPr (encoded by ptsH) and from HPr to one of the sugar specificphosphoryl carrier enzymes II (EII), which usually consist of three(EIIA, EIIB, EIIC) or sometimes four (EIID) domains. From EIIC, whichis located at the inner membrane, the phosphoryl group is transferredto the incoming carbohydrate. In addition to these reactions of uptakeand concomitant phosphorylation of the substrate, the components ofthe C-PTS play an important regulatory role [1–4]. In E. coli especiallyEIIAGlc (encoded by crr) participates in signal transduction dependingon its phosphorylation state [1,2].

Besides the C-PTS in E. coli as well as in many other gram-negativebacteria a second PTS, denoted as the nitrogen-PTS (PTSNtr) [5–7] hasbeen identified by sequence analyses. Similar to the C-PTS a phosphategroup is transferred from PEP to EINtr (encoded by ptsP), NPr (encodedby ptsO/npr) and EIIANtr (encoded by ptsN) (Fig. 1). While the C-PTSworks both in theuptake of different substrates and regulation of carbon

49 3916110510.(K. Bettenbrock).

ights reserved.

metabolism, the PTSNtr seems to have only regulatory functions. In spiteof the location of ptsN and ptsO in one operon together with rpoNencoding the sigma factor 54 (RpoN) (Fig. 1), which is linked to nitrogenstarvation, a clear role of the PTSNtr in nitrogen metabolism or controlunder nitrogen starvation was doubted [8]. However, in a recent publi-cation an effect of α-ketoglutarate and glutamine, both effectors of theNtr response [9–11], on the autophosphorylation activity of EINtr wasdemonstrated [12]. In E. coli several other roles of the PTSNtr e.g. in theregulation of K+ uptake [13,14] and hence in sigma factor selectivity[15], in maintaining cell integrity [5,6,16,17] or in controlling ion fluxes[18] have been discussed. The PTSNtr of Pseudomonas putida has beenreported to be responsible in directing fluxes in the central metabolism[19,20] but also to be active in the production of the overflowmetabolitePHA [21] as well as in several other functions [22–27].

Supported by the observation that in E. coli the inactivation of ptsNinfluenced the growth behavior on different carbon and nitrogensources [6] and that the PTSNtr can exchange phosphate groups withthe C-PTS in vivo [28,29], a function of the PTSNtr in coordinating carbonand nitrogen metabolism seems probable [1,6,30,31]. In order to inves-tigate such a possible regulatory effect of the PTSNtr on central metabo-lism, we analyzed in a systematic approach the growth behavior ofPTSNtr mutants on different carbohydrates (glucose, fructose, N-acetyl-glucosamine (NAG), succinate). Furthermore, we investigated thephosphorylation level of EIIANtr as well as of EIIAGlc and studied gene

rppH

EINtr

rpoN yhbH ptsN yhbJ ptsO

RpoN/σ54 Hpf EIIANtr YhbJ NPr

ptsPrppH rpoN yhbH ptsN yhbJ ptsO

EINtr NPr EIIANtrPEPP

ptsP

EINtr NPr EIIANtrPyruvatP P

Fig. 1. The nitrogen-PTS of E. coli. The genes ptsN (coding for EIIANtr) and ptsO (coding for NPr) are localized in one operon together with rpoN. In the nitrogen-PTS a phosphate from PEP isconferred from EINtr to NPr and from NPr to EIIANtr [6,7].

2880 S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

expression during growth with glucose and succinate. We could showthat mutants lacking a functional EIIANtr displayed a reduced growthrate coupled to an increased acetate production and during growth onglucose in these mutants a downregulation of genes encoding proteinsof the citric acid cycle (TCA) occurred. Mutation of EINtr as well as NPrresulted in a decreased phosphorylation level of EIIANtr while in theparent strain under the conditions tested mainly phosphorylatedEIIANtr could be detected. Metabolic flux analyses of glucose-growncells indicated an increased flux from pyruvate to acetyl-CoA in ptsNmutants. This effect might be due to a missing inhibitory interactionof EIIANtr with PDH. Furthermore, we observed a different phenotypein a mutant lacking functional EIIANtr and NPr than in a mutant lackingonly EIIANtr on the gluconeogenic substrate succinate. While the ΔptsNstrain accumulated mutations in rpoS, the ΔptsN-O strain did not,hinting to a function of NPr for growth on succinate.

2. Material & methods

2.1. Bacterial strains, media and growth conditions

Strains used in this study were LJ110 [32,33], a derivative of E. coliW3110, harboring a frameshift mutation in ilvG, and isogenic mutantsof LJ110 deleted in ptsP, ptsO, ptsN (twovariants), ptsN-yhbJ-ptsO (calledptsN-O) or ptsP plus ptsN-O or in rpoS (Suppl. Table 1). For molecularmethods strains were grown in LB0 medium at 37 °C. Antibiotics sup-plied to the medium were at the following concentrations: kanamycinand chloramphenicol 25 μg/mL; ampicillin (sodium salt) 100 μg/mL;and tetracycline 10 μg/mL.

Growth assays were performed in phosphate-buffered minimalmedium (MM) containing different single carbon sources, which weresterilized by filtration and added to a final concentration of 2 g/L.All experiments were performed in shake flasks, the volumes of theshake flasks exceeding the culture volume at least 10 times. The cul-tures were incubated at 37 °C under vigorous shaking (250 rpm).Growth was monitored by measuring the absorbance at 420 nm inan Ultrospec3000 (Amersham Bioscience). To calculate dry cellweight [g/L] from optical density a factor of 0.312 g/L per OD420

was assumed. MM supplied with 0.2% carbon source was inoculated1:100 with a stationary LB0 culture of the relevant strain (firstculture/preculture). For measuring growth rate, acetate production,mRNA expression, metabolic fluxes and PDH activity, stationaryprecultures were washed in fresh carbon-free MM and fresh MMwith the same carbon source used in the preculture was inoculatedto 5 ∗ 107 cells/mL (second culture).

2.2. Construction of gene deletions and plasmids

Gene deletions were constructed as described by [34]. The com-plete genes were replaced by a chloramphenicol resistance cassette

using pKD3 or by a kanamycin resistance cassette using pKD4 astemplate. Deletion primers contained P1 or P2 priming site in com-bination with 36–40 bp homologous to the region immediately up-stream and downstream of the deleted region. The resistancecassette was eliminated afterwards using pCP20 as described by[34].

For the construction of the plasmid pSJ1 containing ptsN withan C-terminal His6-Tag we used the low copy vector pCS26 [35]containing the constitutive scrK promoter PscrK [36]. ptsN wasamplified by PCR using sj15 providing a synthetic RBS and js16providing a C-terminal His6-Tag (Suppl. Table 2). The DNA fragmentwas cloned as a BamHI-NotI fragment into pCS26PscrK therebydeleting the luxCDABE genes and putting ptsN under the controlof the PscrK.

JS21, a derivate of LJ110 carrying the ptsNH73A allele (encodingEIIANtr lacking the phosphorylation site histidine 73) instead ofptsN, was constructed by gene gorging [37]. RDM (rich definedmedium) [38] was modified and contained per 100 mL: 10 mL 10×MOPS (400 mM MOPS, 500 mM NaCl, 100 mM MgCl2, pH 7.5), 1 mL132 mM K2HPO4, 10 mL 10×-His/-Leu/-Trp/-Ura-dropout-solution(Sigma), 4 mL 25× supplement solution (42,25 mg thiamine-HCl,525 mg histidine, 260 mg tryptophan, 1,32 g leucine per 500 mLH2O) and 2 mL 20% glucose solution. As donor plasmid pSJ19, a derivateof the high copy vector pGEM®Teasy (Promega) containing ptsNH73Aflanked by restriction sites forMeganuclease I-SceI, was used. The insertwas amplified from plasmid pSJ6, containing ptsNH73A, using oligonu-cleotides sj22 and js23. The insert and the vector were ligated using the3′T-overhang of the vector.

For the construction of plasmid pSJ6 encoding ptsNH73A, weused the high copy vector pGEM®Teasy (Promega). ptsNH73A wasamplified by megaprimer PCR [39] using first oligonucleotides sj19and js20 for amplifying the megaprimer, which was used secondlytogether with the oligonucleotide js18 for amplifying ptsNH73A. Astemplatewe used plasmid pSJ2, a derivate of pGEM®Teasy, contain-ing ptsN with an N-terminal His6-Tag PCR-amplified by sj17/js18.Sequences of the different oligonucleotides used are supplied inSuppl. Table 2.

2.3. Measurements of metabolite concentrations

Metabolite concentrations in supernatant were monitored by takingculture samples and removing cells quickly by centrifugation at 4 °C.Measurements were performed enzymatically with the test kits fromR-Biopharm. Production and consumption rates [g/gdcw] were calculatedusing samples taken from the exponential growth phase between OD420

of 0.5 to 1.5. During the exponential phase constant production andconsumption rate were obtained, these values are given in the figuresor tables.

2881S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

2.4. Analysis of gene expression by Real-time RT PCR

To study gene expression mRNAs from at least two independentexperiments were analyzed. From each culture two samples weretaken at different time points in the exponential phase. For immedi-ate stabilization of mRNA two volumes of RNAprotect bacteriaReagent (Qiagen) were added directly to one volume of bacterialculture. Samples were pelleted and stored at −80 °C. Total RNAwas purified from 5 ∗ 108 cells by using Master Pure RNA Purifica-tion Kit (Epicentre) as recommended by the manufacturer. cDNAsynthesis was performed by using RevertAid™ H Minus M-MuLVReverse Transcriptase (Fermentas). 3000 ng total RNA, distributedto six preparations, was reverse transcribed with 600 U ReverseTranscriptase and 600 ng Random Hexamer Primer (Fermentas) at42 °C for 1 h. Afterwards three preparations were pooled and usedfor real-time RT PCR.

Expression of mRNA was quantified by real-time RT-PCR (RT-PCR)in a Rotor-Gene 6000 (Corbett life science) using MESA GREEN qPCRMasterMix Plus for SYBR Assay No ROX (Eurogentec) according to theprotocol supplied. Oligonucleotides were designed by Lasergene PrimerSelect Software (DNAStar) with the following parameters: oligonucleo-tide length was between 18 bp and 22 bp and the amplicon length of90 bp to 180 bp. Oligonucleotides are listed in Suppl. Table 3. PCR cy-cling conditions were 95 °C for 10 min, 40 cycles at 95 °C for 15 s and60 °C for 1 min. Each cDNA pool was measured in duplicate.

For analysiswe exerted themethod of relative quantitative real-timeRT-PCR with efficiency correlation [40,41] using the software qBase(Biogazelle). Geometric mean of transcription levels from recA andrpoDwas used as normalization factor andmRNA level of LJ110 as stan-dard condition (relative gene expression of L110 is hence one for allanalyzed genes).

2.5. Determination of metabolic fluxes

Metabolic fluxes on glucose were analyzed as described beforeusing metabolic flux ratio analysis based on 13C-labeling experi-ments, gas chromatography–mass spectrometry analysis and proba-bilistic equations that relate mass distributions in proteinogenicamino acids to pathway activity [42–45]. To acquire sufficient datatwo sets of experiments were performed: one with the entire amountof glucose being the [1-13C] labeled isotopomer and one where a mix-ture of 20% [U-13C]-glucose and 80% unlabeled glucose was used. Thetotal glucose concentration in the shake flasks was 3 g/L and cellswere incubated in MM on a shaker at 37 °C until mid-exponentialphase. About 5 mg dry cell weight was sampled by centrifugation at4 °C. Further analysis was performed as described by [43]. To comparefluxes of different strains, fluxes of each strain were related to glucoseuptake rate (mmol/gdcw*h). To verify whether ptsN deficient strainsshow stochastically elevated or lower fluxes compared to strains con-taining ptsN we used Whitney U statistics [46]. An influence of a dele-tion of ptsN on metabolic fluxes was called significant at a significancelevel α = 5% for one side test.

2.6. Analysis of EIIAGlc phosphorylation state

The EIIAGlc phosphorylation level was analyzed by sodium dodecylsulfate-polyacrylamide gel electrophoresis and Western blottingessentially as described [36,47]. Protein precipitation was carriedout at −20 °C overnight. Detection was performed as described by[36].

2.7. Analysis of EIIANtr phosphorylation state

To determine the phosphorylation level of EIIANtr amethod based onseparation of the phosphorylated and unphosphorylated EIIANtr vianative PAGEwas applied. 5 ∗ 108 cells were harvested by centrifugation

at 4 °C for 2 min at 16,000 g. The pellet was immediately frozen at−20 °C. Prior to loading on the gel, the pellet was resuspended in100 μL non-denaturing sample buffer containing lysozyme (5 mM Tris–HCl, 40 mM glycine, 10% glycerol, 2 mg/mL lysozyme, pH 8.9) and incu-bated on ice for 15 min. Afterwards the sample was centrifuged for10 min at 16,000 g and 4 °C. 0.0002% bromphenolblue was added tothe supernatant and an aliquot was loaded on the gel. The gel contained2,16 mL acrylamide (30%, mix 29:1), 1,44 mL 5× running buffer(125 mM Tris–HCl, 1 M glycine, pH 8.9), 3,36 mL H2O, 15 μL TEMEDand 36 μL APS (10%). After loading of the samples the gel was run at12.5 mM at 4 °C in 1× running buffer.

After completion of the run the proteins were transferred to a PVDFmembrane via semi-dry blotting at 15 V for 90 min. Prior to blotting thegel was equilibrated twice in transfer buffer (192 mM glycine, 25 mMTris, 29% methanol, 0.0375% SDS) for 15 min.

Themembrane was washed two times for 10 min in washing buffer(1× PBS, 0.01% Tween20). Afterwards themembrane was blockedwith3% skimmedmilk powder, washed three timeswithwashing buffer andthen incubated with rabbit serum raised against His6-EIIANtr for 1 h.Detection was performed with horseradish-peroxidase coupled goat-anti-rabbit antibodies and the “Supersignal West Femto MaximumSensitivity Substrate” (Pierce). Signalswere detectedwith a CCD camerathe two EIIANtr bands were quantified relative to each other.

2.8. Determination of PDH activity

For measuring PDH activity cells were cultured in minimal mediumwith glucose. In the exponential growth phase, at an OD420 = 1 ± 0.1525 mL strain culture was harvested by centrifugation. The pellet waswashed two times with 10 mL of extraction buffer (100 mM Tris–HCl,10 mM MgCl2, pH 7.2) and afterwards resuspended in 1.5 mL of thesame buffer. The probe was sonicated on ice and debris was collectedby centrifugation. The protein concentration of the supernatant wasmeasured by Bradford test [48]. PDH activity was measured in accor-dance to [49]. The reaction mixture (1 mL) contained 100 mM Tris–HCl(pH 7,2), 10 mM MgCl2, 1 mM TPP, 5 mM pyruvate, 3 mM L-cysteine,2 mM NAD, 100 μM coenzymeA and 100 μL crude extract. The reactionmixturewas preincubated for 3 min at 37 °C and the reactionwas startedby adding coenzymeA. The specific activity wasmonitored spectrophoto-metrically by the increase in 340 nm absorbance at 37 °C due to NADH2

production. Enzymatic activity was calculated as micromoles of sub-strate converted per milligram of protein per minute (U/mg protein).

2.9. Determination of rpoS mutants

For expressing catalase and producing glycogen, E. coli cellsneed an active rpoS-allele [50,51]. Therefore a mutation of rpoScould be verified by qualitative measuring of catalase activity andglycogen production in cells cultured on LB0 agar at 37 °C untilthey reached stationary phase (16–18 h). Catalase activity wastested by incubating cells with a hydrogen peroxide solution(6%). While rpoS+ cells bubble immediately and intensive, rpoS−

cells bubble time-delayed and slight. Glycogen production was tested byincubating cells with an iodine–potassium iodine-solution (0.01 M I2and 0.03 M KI). While rpoS+ cells get a brown color, rpoS− cellsremain unstained.

3. Results

3.1. Determination of growth rates and acetate production in PTSNtr deletionmutants

To get information about a possible function of the PTSNtr in reg-ulating carbon metabolism, we analyzed the growth of E. coli LJ110, aderivative of W3110, harboring a frameshift mutation in ilvG, and ofisogenic ΔptsP, ΔptsO, ΔptsN and ΔptsN-O strains on defined single

2882 S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

carbon sources, comprising PTS substrates (glucose, fructose, N-acetylglucosamine (NAG)) as well as the non-PTS substrate succi-nate. Glucose, which is taken up by the glucose-PTS, represents thepreferred growth substrate of E. coli. Fructose was used because thefructose-PTS contains the diphosphoryltransfer protein (DTP) whoseEIIA domain shows high homology to EIIANtr [6,31]. The PTS-substrateNAG represents a C-source as well as a N-source. The organic acid suc-cinate feeds directly into the TCA and hence complements the abovementioned glycolytic substrates.

As expected, for all strains the growth rates differed with the carbonsource (Fig. 2). Growth on glucose was quickest for all strains, whilegrowth on succinate was slowest. The mutant strains showed a similargrowth pattern on all glycolytic substrates tested. While the ΔptsP andthe ΔptsO strain always behaved like the parent strain, the ΔptsN andthe ΔptsN-O strain grew significantly slower. Uptake rates for glucosewere determined for the different PTSNtr mutants (Suppl. Table 6).They show in accordance with the reduced growth rates reducedglucose uptake rates for the ΔptsN and the ΔptsN-O mutants. Thus thedata show that a deletion of ptsN affects the growth behavior and indi-cate that for the growth promoting effect of EIIANtr no phosphorylationthrough EINtr and NPr is needed.

On succinate, too, the ΔptsO and the ΔptsP strain grew similar to theparent strain but on this substrate differences between the ΔptsN andthe ΔptsN-O strain became visible. The ΔptsN strain grew extremelyslow (μ = 0.11 ± 0.01) after transfer from LB0 into minimal medium(MM) with succinate but growth was relatively fast (μ = 0.30 ±0.02) after cells from stationary phase of this MM culture (first culture)were inoculated into fresh MM with succinate (second culture). Thisdifference in growth rate between the first and the second culturehints to an accumulation of secondary mutations in the ΔptsN strainduring growth on succinate (see below). The ΔptsN-O strain on thecontrary grew significantly slower than the parent strain but in thiscase the difference between the first (μ = 0.17 ± 0.00) and the secondMM culture (μ = 0.23 ± 0.02) was less pronounced. Wildtype growthof both theΔptsN and theΔptsN-O strain on all substrates testedwas re-stored by complementing the mutants with a plasmid pSJ1 expressingptsN (not shown).

Growth of E. coli under carbon excess conditions is characterized bythe formation and excretion of acetate in the overflowmetabolism [52].The amount of acetate excretion differs with the carbon source. It isknown that the PTSNtr has an influence on PHA production in P. putida[21] and affects polyhydroxybutyrate levels in Azotobacter vinelandii[53], both products of carbon overflow.We thus determined extracellu-lar acetate yields in the exponential growth phase for the parent strainas well as for the deletion strains (Fig. 2).

As expected, acetate yields differed in dependence of the carbonsource. All strains produced the highest amount of acetate on NAG,

Fig. 2. Growth rate (left) and acetate production (right) of LJ110 and isogenic ptsP, ptsO, ptsN anminimalmediumcultures growingwith the carbon source indicated (second culture). Acetate pthe second culture as described in Materials & methods.

which most probably results from the conversion of NAG-P to D-glu-cosamine-6-phosphate and acetate catalyzed by NagA. But while nosignificant differences in acetate formation were detectable for theΔptsP and the ΔptsO strain, acetate yields of the ΔptsN and the ΔptsN-Ostrain were elevated on the glycolytic substrates (Fig. 2), reflectinggrowth behavior. The strains produced acetate also during growthon succinate but in this case the ΔptsN strain (2nd MM culture) aswell as the ΔptsP and ΔptsO strain showed an acetate yield comparableto LJ110 while that of the ΔptsN-O strain was again significantly ele-vated. This observation correlates with the growth data, which showeda different growth behavior of these strains on succinate. As observedfor growth rate, expressing EIIANtr from plasmid pSJ1 restored acetateproduction in both ptsNmutants (not shown).

The results of studying acetate formation hence confirmed the dataof the growth survey. On each carbon source tested, slower growth cor-related with a higher production of acetate. Acetate production by car-bon overflow is linked to the Pta-AckA pathway that produces acetateand ATP from acetyl-CoA. In addition, CoA-SH is released. To distinguishwhether enhanced acetate formation is the reason for the observedslow growth phenotype of the ΔptsN and the ΔptsN-O strain, an addi-tional deletion of the pta-ackA genes was introduced into thesemutantsand into LJ110. Although all strains showed drastically reduced acetateproduction, the Δ(ackA-pta) mutation had no effect on the observedgrowth behavior on glucose and succinate (not shown). Hence, acetateformation most probably is not the cause for the slow growth rates ofthe ptsN mutants but a consequence of changed fluxes in the centralmetabolism. This hints to a role of the PTSNtr or more precisely of EIIANtr

in controlling fluxes in central metabolism especially at the acetyl-CoAnode and the TCA cycle.

Summarizing, the growth data suggest that EIIANtr promotes growthon all carbon sources. Inactivation of EINtr and NPr had no effect ongrowth, indicating that phosphorylation of EIIANtr by these proteins isnot important for its function under the conditions tested. However,the data hint to a function of NPr during growth on the gluconeogenicsubstrate succinate, an effect which is only observable in a ΔptsNbackground.

3.2. Determination of EIIANtr and EIIAGlc phosphorylation state

Assuming thephosphorelay EINtr → NPr → EIIANtrmutations in ptsPand ptsO are supposed to change the phosphorylation state of EIIANtr.We analyzed the EIIANtr phosphorylation state by using native gel elec-trophoresis coupled to Western blotting with antibodies raised againstEIIANtr. In the development of this method a strain was constructedwhich expresses EIIANtr(H73A), harboring a histidine to alanine ex-change at position 73, the phosphorylation site. This strain was used asa non-phosphorylated standard. EIIANtr phosphorylation levels were

d ptsN-O deletion strains on different carbohydrates. Growth rates were determined fromroductionwas determined from four tofive time points in the exponential growth phase of

LJ1

10

ptsN

H73

A

ΔptsOGlc

ΔptsPNAG

ΔptsPGl

EIIANtr

EIIANtr P

sNH

73A

LJ1

10

EIIANtr

pts

-

EIIANtr P-

NAG Glc

Fig. 3. In vivo phosphorylation state of LJ110 and of the ΔptsO and ΔptsP strains duringgrowth with glucose or NAG, respectively. Phosphorylation levels were determined byrelating the two bands representing phosphorylated and dephosphorylated EIIANtr.LJ110 ptsNH73A expressing the EIIANtr(H73A) variant shows only the dephosphorylatedband. On glucose for LJ110 almost no dephosphorylated EIIANtr was detectable, whilethe mutants displayed phosphorylated as well as dephosphorylated EIIANtr. Multiplelanes for the same strain and substrate represent different samples taken during a timecourse experiment.

2883S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

determined from the growth experiments described above (Fig. 3,Table 1). Confirming the results of Bahr et al. [54], on glucose mainlyphosphorylated EIIANtr was observed for the parent strain LJ110, whilein the ΔptsP and the ΔptsO strain the phosphorylation state of EIIANtr

was significantly reduced. A lower phosphorylation level was observedin the ΔptsP strain lacking EINtr (16%) than in the ΔptsO strain lackingNPr (57%). While in LJ110 EIIANtr was mainly in its phosphorylatedstate on all growth substrates tested, in the ΔptsP and the ΔptsO straina carbon source dependent phosphorylation became visible. A carbonsource dependent phosphorylation also applies for EIIAGlc [36]. Thedata hint to a crosstalk between the PTSNtr and the C-PTS, that wasalready reported by [6,7,28]. This crosstalk becomes obvious in mutantswith defective PTSNtr phosphorelay.

Via cross talk mutations in the PTSNtr might influence the phosphor-ylation state of the C-PTS. Especially EIIAGlc (encoded by crr) from theC-PTS, participates in signal transduction and regulation of carbonmetabolism in dependence of its phosphorylation state [1,47,55–63].It thus seems possible that deletions in the PTSNtr influence growth be-havior on different substrates indirectly by changing the phosphoryla-tion state of EIIAGlc. We analyzed EIIAGlc-P levels in the exponentialgrowth phase in the parent as well as in the PTSNtr mutant strains. Thephosphorylation state of EIIAGlc varied depending on the kind of sub-strate as described before [36] but it did not vary in the mutant strains(Table 1). The observed almost complete dephosphorylation of EIIAGlc

Table 1Phosphorylation state of EIIANtr and EIIAGlc during growth onminimal medium (second culturefraction of EIIANtr or EIIAGlc respectively given in %. The data are average data of at least 2 indepeexperiment.

EIIANtr-P [%]

Strain\C-source LJ110 ΔptsP ΔptsO

Glc 95 ± 9 16 ± 3 57 ± 13Fru 97 ± 8 35 ± 13 67 ± 8NAG 96 ± 12 15 ± 3 44 ± 7Suc 96 ± 7 82 ± 11 87 ± 9

in the ptsN mutants on glucose is noteworthy. Previously, a couplingof growth rate and EIIAGlc phosphorylation level has been shown[36,64]. Based on this, a higher level of EIIAGlc phosphorylation wasexpected for the slower growingptsNmutants. The fact the EIIAGlc phos-phorylation in the case of the ptsN mutants no longer is coupled togrowth rate hints to an influence of EIIANtr on the C-PTS either directlyor indirectly by influencing the PEP to pyruvate ratio.

3.3. Determination of transcription levels of TCA enzymes via real-timeRT-PCR

Our results from the analysis of growth behavior and acetate forma-tion in the PTSNtr mutants indicated a regulation of carbon metabolismthrough EIIANtr and possibly NPr.

It is known, that during growth on glucose higher acetate formationcorrelates negatively with transcription of TCA genes [52]. We thus in-vestigated transcription of the genes gltA, sucB, sdhD andmdh encodingcitrate synthetase, α-ketoglutarate dehydrogenase, succinate dehydro-genase and malate dehydrogenase. mRNA levels were determined forLJ110, as well as for the ΔptsO, the ΔptsN and the ΔptsN-O strain viaquantitative real-time RT-PCR. To obtain further information about theregulation of centralmetabolismwe also observed pykF and pykA (pyru-vate kinase I and II), aceE (pyruvate dehydrogenase), aceA (isocitratelyase) and ackA (acetate kinase) expression. recA and rpoD were takenas housekeeping genes.

At first we analyzed gene expression during exponential growth onglucose (Fig. 4, Suppl. Table 4). In linewith the acetate production, tran-scription of the TCA genes gltA, sucB and sdhD and the glyoxalate shuntgene aceAwas about twofold downregulated in the ΔptsN and ΔptsN-O,but not in the ΔptsO strain. In addition, the expression of the TCA genemdhwas also downregulated in theΔptsN andΔptsN-O strain, althoughit was less pronounced than the downregulation of the other TCA genes.Expression of pykF, pykA and ackA was similar in all strains. Notably,expression of aceE, encoding a subunit of PDH-complex,was unchanged.A similar expression pattern for a ptsNmutant has also been observed by[15], although its reported especially low growth rate on glucose atpotassium concentrations comparable to those used in this study didnot match our results.

As expected, the transcription patterns on succinate differed fromthose on glucose (Suppl. Table 4). Expression of most TCA genes wasupregulated in LJ110 compared to growth on glucose. Previously, lacZ-fusion studies showed an upregulation of the TCA cycle genes gltA[65], fumA and fumC [66], mdh [67], sdhCDAB [68], sucABCD [69] andicd [70] on succinate. Additionally, [71] showed increased activity ofseveral Krebs cycle enzymes in malate containing complex mediumcompared to glucose containing complexmedium.Malate is an oxidizedTCA intermediate as is succinate. Microarray data showed an at least3-fold upregulation of aceA on succinate compared to glucose [72]confirming the results obtained in this study. Contradicting previousresults no difference inmdh expression on succinate and glucose wasobserved. The twofold downregulation of pykF expression verifiedresults from lacZ fusion studies [73].

The transcription patterns of the different mutant strains on succi-nate (Fig. 5, Suppl. Table 5) reflected their growth behavior in the

) supplemented with different carbon sources. The values given show the phosphorylatedndent experiments and at least four time points during the exponential growth phase per

EIIAGlc-P [%]

LJ110 ΔptsP ΔptsO ΔptsN

6 ± 3 4 ± 1 8 ± 2 9 ± 329 ± 8 34 ± 8 37 ± 8 43 ± 89 ± 3 5 ± 2 6 ± 1 5 ± 1

82 ± 5 85 ± 4 80 ± 8 76 ± 8

1

0.5

-0.5

0

-1

expr

esio

n le

vel

-2

-1.5

pykF sucBpykA ackA aceE gltA aceA sdhC mdhpykF pykA ackA aceE gltA aceA sucB sdhC mdh

ptsO ptsN ptsN-O

Fig. 4.mRNA expression of selected genes in the ΔptsO, ΔptsN and ΔptsN-O strains during exponential growth on glucose. Expression was scaled to LJ110 growing on glucose. As house-keeping genes rpoD and recAwere used.

2884 S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

second MM culture. While the ΔptsO and the ΔptsN strain (2nd MMculture) showed a transcription pattern similar to the parent strain,the ΔptsN-O strain differed. The expression of pykF, encoding pyruvatekinase I, was twofold upregulated in this strain, the expression of aceA,encoding isocitrate lyase, was twofold downregulated. Because of thedifferent growth behavior of the ΔptsN strain in the first and secondminimal medium cultures on succinate, the transcription pattern deter-mined might result from unknown mutations accumulating in thisstrain and allowing growth similar to the parent strain. Therefore,

1.5

1

0

0.5

-0.5

-1.5

-1

ptsO

log 2

rel.

expr

essi

on

pykF pykA aceE ackA gltpykF pykA aceE ackA glt

Fig. 5. Expression of selected genes in theΔptsO,ΔptsN andΔptsN-O strain during exponential ggenes rpoD and recAwere used.

during growth on succinate the presence of NPr seems to lead tosecondary mutations when EIIANtr is inactive. This hints to a functionof NPr during growth with succinate. Interestingly this functionbecame visible only when NPr was present and EIIANtr was absent.

3.4. Metabolic fluxes on glucose are altered in ptsN mutants

Because the growth data as well as the real-time data indicated aninfluence of EIIANtr on carbon metabolism, we measured metabolic

ptsN ptsN-O

A aceA sucB sdhC mdhA aceA sucB sdhC mdh

rowth on succinate. Expressionwas scaled to LJ110 growing on succinate. As housekeeping

2885S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

fluxes in themutants and the parent strain growing on glucose. BesidesLJ110 and the ΔptsP, ΔptsO, ΔptsN and ΔptsN-O strains also a ΔptsNkeio

strain, a ΔptsN strain constructed following the KEIO collection [74]and a strain combining the ptsP and the ptsN-O deletion, were analyzed.Both strains behaved similar as the ΔptsN strain in growth rate andacetate production on glucose (not shown). Average fluxes of two tech-nical replicates of each strain are shown in Suppl. Table 6.

The glucose consumption rate per biomass was similar for all strainstested, while the acetate production rate was significantly higher inall ptsN deficient strains. Normalized to glucose uptake rate thefluxes through the TCA (from oxaloacetate to α-ketoglutarate, fromα-ketoglutarate to malate and from malate to oxaloacetate (Mdh)were downregulated in the ΔptsN and ΔptsN-O strain as well as inall other ptsN deleted strains, confirming the data showing loweredexpression of TCA genes in the ΔptsN and ΔptsN-O strains. Asexpected from the measurement of extracellular acetate yield, theflux from acetyl-CoA to acetate (Pta-AckA) was upregulated in allptsN mutants. In addition, the flux from pyruvate to acetyl-CoA(PDH) was elevated and flux from PEP to oxaloacetate (Ppc) wasreduced in all ptsN mutants. Tested by Whitney U statistics thesechanged fluxes resulted from the deletion of ptsN at a significancelevel of α = 5%. For all other fluxes no significant changes could bedetected between ptsN+ and ptsN− strains. A flux diagram comparingfluxes of LJ110 and of its ΔptsN derivative is shown in Fig. 6.

To further confirm the data from the 13C flux analysis, we measuredPDH activity in cell extracts of LJ110 and the ΔptsN strain. For LJ110 aPDH activity of 0.35 ± 0.05 μmol/mg/min was determined comparedto 0.55 ± 0.14 μmol/mg/min of theΔptsN strain. The data of the enzy-matic assay hence confirm the data from the flux analysis indicatinga higher PDH activity in the absence of ptsN. EIIANtr hence mightrepresses PDH activity by protein–protein interaction as has been

Fig. 6. Relative fluxes for LJ110 and an isogenic ΔptsN strain during aerobic growth on glucose. Gglucose flux to 100. Flux arrows are drawn in proportion to the substrate uptake for each strdescribed in Section 2.5. The presented fluxeswere obtained 13C-constrained flux analysis usingcatalyzing theflux. TCA1 is a flux comprising the reactions from oxaloacetic acid and acetyl-CoADH). TCA2 comprises the reactions from α-ketoglutarate to malate catalyzed by SucAB (2-o(fumarase).

shown for P. putida [20], supported by the fact that no changes ingene expression of aceE encoding a PDH subunit were detected. Adeletion of ptsN results in an increased activity of the PDH, which isobserved in a higher flux from pyruvate to acetyl-CoA. As the in-creased amount of acetyl-CoA cannot be channeled into the TCAcompletely due to the downregulation of most genes encoding TCAenzymes, this results in increased overflow to acetate. Why andhow downregulation of genes encoding TCA enzymes is accom-plished is not known. Based on our experiments it cannot be excludedthat other fluxes at the pyruvate/PEP/acetyl-CoA node or in the TCAare also influenced directly or indirectly by the PTSNtr e.g. fluxes to orfrom lipid synthesis.

3.5. ptsN mutants accumulate rpoS mutations during growth withsuccinate

As already mentioned above the ΔptsN strain showed no constantgrowth behavior during growth with succinate. It adapted very slowlyfrom growth in LB0 to growth in minimal medium with succinate.Further tests (culturing of cells fromdifferent growth phases inMMme-diumwith an intermediate culture in LB0) indicated that this is not dueto a slow adaptation but due to the accumulation ofmutations. ptsNmu-tants accumulate higher intracellular amounts of K+ [15] which favorsbinding of σS instead of σ70 to RNA polymerase. Based on this observa-tion we investigated σS in the ptsN mutants during growth with succi-nate. A mutation in rpoS encoding σS is coupled to a lack of glycogenaccumulation as well as to a loss of catalase [50,51]. We hence grewcells of LJ110 aswell as of theΔptsN andΔptsN-O strain in LB0 overnightand inoculated them inminimalmediumwith 0.4% succinate or glucoserespectively to an OD420 of 0.025–0.045. The cultures were incubated at37 °C and cells from early and late growth phase were plated on LB0.

rowth rates are given as well as glucose uptake rates thatwere used to scale the incomingain. Values represent the average of two technical replicates. Samples were prepared asthe software FiatFlux [43–45]. The flux names given correspond to the respective enzymetoα-ketoglutarate catalyzed byGltA (Citrate synthase), Acn (aconitase) and Icd (isocitratexoglutarate DH), SucCD (SuccinylCoA synthetase), SdhABCD (Succinate DH) and FumA

2886 S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

Colonies grown on these plates were analyzed with respect to glycogenaccumulation and catalase activity. On succinate a relevant portion ofthe colonies from the ΔptsN strain produced no glycogen and showedno catalase activity (52% at an OD420 0.36 and 79% at an OD420 4.8),indicating the accumulation of secondary mutations in rpoS. In con-trast, for the ΔptsN-O strain as well as for LJ110 no rpoS mutationsaccumulated during growth on succinate. In addition, no rpoSmutants accumulated during growth with glucose in LJ110 as wellas in the ΔptsN and ΔptsN-O strain.

Although rpoSmutations accumulated in the ΔptsN strain, it cannotbe ruled out that further mutations are necessary in order to allow forfast growth of the ΔptsN strain on succinate. To asses this, we deletedrpoS in the parent strain LJ110 as well as in the ΔptsN and in theΔptsN-0 mutants and analyzed the growth behavior of the corre-sponding mutants on MM glucose and on MM succinate. The deletionof rpoS in LJ110 had no significant influence on growth (Fig. 7). Inactiva-tion of rpoS led to a significantly shortened lag phase of theΔptsN strainon glucose (not shown) and also on succinate (Fig. 7). Against expecta-tion also the lag phase of the ΔptsN-O strain was reduced. Notably,growth rate and acetate production in the second culture were almostunaffected by the ΔrpoS mutation. These data confirm an inhibitingeffect of RpoS on the adaption of ptsNmutants from complex tominimalmedium. But this effect obviously is limited to the adaptation phase andis not determining specific growth rate or acetate production afteradaptation.

Taken together these data hint to an important function of EIIANtr

during growth with succinate, which is dependent on the presence ofNPr. Obviously, the negative effect of a ptsNmutation for growth on suc-cinate can be released by amutation in rpoS, most probably by effectinggene expression during the adaptation phase.

4. Discussion

To investigate a possible function of the PTSNtr in controlling fluxesin the central metabolism, we performed quantitative growth assaysof E. coli LJ110 and of isogenic deletion mutants lacking different PTSNtr

proteins.On all substrates tested, the ΔptsN and ΔptsN-O strains showed

lower growth rates coupled to increased acetate production indicating

10

1

0.1

OD

420

0.010 5 10

tim

Fig. 7.Growth curve of theparent strain LJ110 and of theΔptsN andΔptsN-Omutants aswell as odata for the first minimal medium culture. It can be observed that the rpoSmutation causes a

a growth promoting function of EIIANtr. In contrast, the ΔptsP andΔptsO strain behaved similar to the parent strain although the absenceof EINtr or NPr resulted in a reduced phosphorylation level of EIIANtr

(Table 1). The EIIANtr phosphorylation state is obviously not importantfor its growth promoting function. This was unexpected as the activityof PTS proteins in general and also of EIIANtr has been reported to beinfluenced by their phosphorylation state [13,14,16–18,47,55,57,62].Although the dephosphorylated form of EIIANtr has been implied intothe regulatory actions elucidated so far [13,14,18,19,75], we were notable to detect significant amounts of dephosphorylated EIIANtr in theparent strain. Only in the ΔptsP and ΔptsO strains a significant fractionof dephosphorylated EIIANtr was observed but obviously this did nothave an influence on growth. The phosphorylation levels determinedhere are in agreement with those determined by Bahr et al. [54]. Theyagain hint to a cross talk between the C-PTS and PTSNtr but it can bedoubted that this cross talk is important under wildtype conditions,based on the fact that phosphorylation levels of EIIAGlc and EIIANtr inthe wildtype are not correlated. Deviating phosphorylation levels ofEIIANtr were reported by Lee et al. [12] who observed mainlyunphosphorylated EIIANtr during growth in minimal medium withglucose and 40 mM ammonia. Lee et al. [12] used an EIIANtr variantwith an amino acid substitution near the phosphorylation site thatmight influence phosphorylation kinetics within the PTSNtr or crosstalk with the C-PTS. Likewise it cannot be excluded that expression ofEIIANtr from a plasmid as performed by Lee et al. [12] or differences inthe strain background are responsible for the deviating results.

On glucose slow growth and enhanced acetate production of theΔptsN and the ΔptsN-O strain were coupled to a reduced expression ofgenes encoding TCA enzymes (Fig. 4). This was confirmed by 13C fluxanalyses showing reduced fluxes through the TCA as well as an elevatedflux from acetyl-CoA to acetate in thesemutants (Fig. 6). In the extractsof aΔptsN strain a higher PDH activitywas detected, supported by an in-creased flux from pyruvate to acetyl-CoA (Pdh) normalized to glucoseuptake rate. In P. putida a direct interaction of EIIANtr with the E1 subunitof the pyruvate dehydrogenase complex (PDH), leading to a down-regulation of PDH activity was detected [20]. It might hence be possiblethat in E. coli EIIANtr regulates the flux from pyruvate to acetyl-CoA byinteracting with PDH, too. This is approved by the observation that inthe ptsN mutants a decoupling of EIIAGlc phosphorylation state from

LJ110

Δ

ΔptsN-O

ΔrpoS

ΔrpoS ΔptsN

15 20 25 30

ΔrpoS ΔptsN-O

e [h]

LJ110

ptsN

f the three strainswith an additionalΔrpoSmutation onMM0.2% succinate. Shown are thefaster adaptation to MM succinate in all strains.

2887S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

growth rate was observed. Previously, it was reported that the phos-phorylation state of EIIAGlc reflects growth rate [36,64]. The EIIAGlc phos-phorylation level is determined by the PEP to pyruvate ratio, which iscoupled to growth rate in E. coli [36,76,77]. Obviously, this relationbetween growth rate and EIIAGlc phosphorylation is disturbed in theptsN mutants. Considering the observation that the phosphorylationlevel of EIIAGlc is the main factor in determining cAMP productionin E. coli and the fact that cAMP.CRP controls many genes encodingnutrient uptake systems as well as e.g. enzymes of the TCA cycle[1,2] the low phosphorylation level in the ptsN mutants might beone reason for the slow growth phenotype.

Lee et al. [12] observed an effect of extracellular ammonium concen-tration as well as of different nitrogen sources on the phosphorylationactivity of EINtr. Similarly EI autophosphorylation is controlled byα-ketoglutarate [78]. α-Ketoglutarate as well as glutamine areimportant metabolites in the control of ammonium assimilation[10,11]. α-Ketoglutarate is an intermediate of the TCA cycle andconnects carbon- and nitrogen-metabolism. Based on this and onthe data presented here it is tempting to speculate that the PTSNtr

is involved in the coordination of C- and N-metabolism by adjustingthe fluxes at the acetyl-CoA node and in the TCA cycle.

In part disaccording data result from growth on succinate. As alreadyreported, there are striking differences in the growth of the ΔptsN andΔptsN-O strains. While the ΔptsN-O strain is able to grow on succinate,theΔptsN strain seems to be able to growonly after acquiringmutationsin rpoS. It is thus tempting to speculate that also NPr has a functionduring growth on succinate. In contrast to the other carbohydratestested growth on succinate requires an active gluconeogenesis andhence partly reversed fluxes in the central metabolism. A regulatoryfunction of NPr in gluconeogenesis seems possible. Because thisfunction is only relevant for growth when EIIANtr is absent, it canbe concluded that EIIANtr modulates the regulatory mechanism ofNPr e.g. by modulating its phosphorylation state. A deletion of ptsNcould lead to an elevated phosphorylation state of NPr and therebyactivate the regulation by NPr resulting in reduced gluconeogenicfluxes. As during growth on succinate fluxes have to be directedfrom oxaloacetate or malate to phosphoenolpyruvate and pyruvateit is tempting to speculate that one of these reactions is controlled.In a ΔptsN strain this inhibitory effect adds to the elevated fluxfrom pyruvate to acetyl-CoA and the reduced TCA flux. As a conse-quence growth on succinate is severely inhibited and secondarymutations accumulate. By inactivating NPr gluconeogenic fluxesare restored.

In viewof the fact that expressionof a highnumber of genes is affectedby a ΔptsNmutation an effect of the PTSNtr on global control is obvious.A possible influence of the PTSNtr on catabolite repression seems prob-able based on the observed low phosphorylation state of EIIAGlc in theslow growing ptsN mutants on glucose. In addition, an effect of EIIANtr

on sigma factor selectivity by influencing the intracellular K+ concen-trations has been reported [15]. RpoS activity has been linked to theactivity of the ArcBA two-component system [79], a system that con-trols a high number of genes especially those of the TCA cycle and therespiratory chain in dependence of the oxygen concentrations [80–82]matching the observed gene expression profiles. An effect of ptsNmuta-tions on the activity of RpoS could also be demonstrated in this study, asduring growth on succinate, the only gluconeogenic substrate tested,rpoSmutations accumulated in the ΔptsNmutant. Growth on succinaterequires fluxes from the TCA in the direction of PEP or pyruvate. This isrealized by either PEP carboxykinase or by malic enzyme and phospho-enolpyruvate synthetase [83]. It is hence tempting to speculate thatexpression of these genes is somehow blocked by an active RpoS. Anactivating influence of the PTSNtr on the enzymes of the pyruvateshuntwas observed in P. putida [19] but as these observations are relatedto growth with glucose and fructose and as the usage of metabolic path-ways in E. coli and P. putida is different, it remains open if the underlyingeffects are comparable.

Summarizing, the studies presented here suggest independent func-tions of EIIANtr and NPr in E. coli. While EIIANtr fulfills an important roleduring aerobic growth in minimal medium with different substrates bycontrollingfluxes at the acetyl-CoAnode and in the TCA cycle, the role ofNPr only becomes visible in a ΔptsN background during growth onsuccinate and might hence be related to gluconeogenesis. Analysis ofthe EIIANtr phosphorylation state indicated that phosphorylation ofEIIANtr is not relevant for its function in controlling metabolic fluxes. Inaddition, cross talk between the PTSNtr and C-PTS was observed in vivoin mutant strains but seems to have no relevance in determining thephosphorylation state of the proteins of the PTSNtr and C-PTS underwildtype conditions.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbamcr.2013.07.011.

References

[1] J. Deutscher, C. Francke, P.W. Postma, How phosphotransferase system-related pro-tein phosphorylation regulates carbohydrate metabolism in bacteria, Microbiol.Mol. Biol. Rev. 70 (2006) 939–1031.

[2] P.W. Postma, J.W. Lengeler, G.R. Jacobson, Phosphoenolpyruvate: carbohydratephosphotransferase systems of bacteria, Microbiol. Rev. 57 (1993) 543–594.

[3] J.W. Lengeler, K. Bettenbrock, R. Lux, Signal-transduction through phosphotransferasesystems, in: A. TorrianiGorini, E. Yagil, S. Silver (Eds.), Phosphate in Microorganisms—Cellular and Molecular Biology, 1994, pp. 182–188.

[4] S. Roseman, N.D. Meadow, Signal transduction by the bacterial phosphotransferasesystem — Diauxie and the Crr Gene (Monod, J. Revisited), J. Biol. Chem. 265 (1990)2993–2996.

[5] K. Pflüger-Grau, B. Görke, Regulatory roles of the bacterial nitrogen-relatedphosphotransferase system, Trends Microbiol. 18 (2010) 205–214.

[6] B.S. Powell, D.L. Court, T. Inada, Y. Nakamura, V. Michotey, X.W. Cui, A. Reizer, M.H.Saier, J. Reizer, Novel proteins of the phosphotransferase system encoded withinthe rpoN operon of Escherichia coli—enzyme IIANtr affects growth on organicnitrogen and the conditional lethality of an erats mutant, J. Biol. Chem. 270(1995) 4822–4839.

[7] R. Rabus, J. Reizer, I. Paulsen, M.H. Saier, Enzyme INtr from Escherichia coli — anovel enzyme of the phosphoenolpyruvate-dependent phosphotransferasesystem exhibiting strict specificity for its phosphoryl acceptor, NPr, J. Biol.Chem. 274 (1999) 26185–26191.

[8] A.J. Ninfa, Unnecessary signaling: poorly named? J. Bacteriol. 193 (2011) 4571–4573.[9] L.J. Reitzer, Ammonia assimilation and the biosynthesis of glutamine, glutamate, as-

partate, asparagine, L-alanine, and D-alanine, in: F.C. Neidhardt, E.C.C. Lin, K. BrooksLow, B. Magasanik, W.S. Reznikoff, M. Riley, M. Schaechter, H.E. Umbarger (Eds.),Escherichia coli and Salmonella: Cellular and Molecular Biology, ASM Press, Wash-ington, DC, 1996, (http://ecosal.org).

[10] A.J. Ninfa, P. Jiang, PII signal transduction proteins: sensors of alpha-ketoglutaratethat regulate nitrogen metabolism, Curr. Opin. Microbiol. 8 (2005) 168–173.

[11] L. Reitzer, B.L. Schneider, Metabolic context and possible physiological themesof σ54-dependent genes in Escherichia coli, Microbiol. Mol. Biol. Rev. 65 (2001)422–444.

[12] C.R. Lee, Y.H. Park, M. Kim, Y.R. Kim, S. Park, A. Peterkofsky, Y.J. Seok, Reciprocalregulation of the autophosphorylation of enzyme I by glutamine and alpha-ketoglutarate in Escherichia coli, Mol. Microbiol. 88 (2013) 473–485.

[13] C.R. Lee, S.H. Cho, M.J. Yoon, A. Peterkofsky, Y.J. Seok, Escherichia coli enzymeIIA(Ntr) regulates the K+ transporter TrkA, Proc. Natl. Acad. Sci. U. S. A. 104 (2007)4124–4129.

[14] D. Lüttmann, R. Heermann, B. Zimmer, A. Hillmann, I.S. Rampp, K. Jung, B. Görke,Stimulation of the potassium sensor KdpD kinase activity by interaction with thephosphotransferase protein IIA(Ntr) in Escherichia coli, Mol. Microbiol. 72 (2009)978–994.

[15] C.R. Lee, S.H. Cho, H.J. Kim, M. Kim, A. Peterkofsky, Y.J. Seok, Potassium mediatesEscherichia coli enzyme IIA(Ntr)-dependent regulation of sigma factor selectivity,Mol. Microbiol. 78 (2010) 1468–1483.

[16] J.D. Hayden, S.E. Ades, The extracytoplasmic stress factor, sigmaE, is required tomaintain cell envelope integrity in Escherichia coli, PLoS One 3 (2008) 13.

[17] H.J. Kim, C.R. Lee, M. Kim, A. Peterkofsky, Y.J. Seok, Dephosphorylated NPr of the ni-trogen PTS regulates lipid A biosynthesis by direct interaction with LpxD, Biochem.Biophys. Res. Commun. 409 (2011) 556–561.

[18] D. Lüttmann, Y. Göpel, B. Görke, The phosphotransferase protein EIIANtr modulatesthe phosphate starvation response through interaction with histidine kinase PhoRin Escherichia coli, Mol. Microbiol. 86 (2012) 96–110.

[19] M. Chavarria, R.J. Kleijn, U. Sauer, K. Pflüger-Grau, V. de Lorenzo, Regulatory tasks ofthe phosphoenolpyruvate-phosphotransferase system of Pseudomonas putida incentral carbon metabolism, MBio 3 (2012).

[20] K. Pflüger-Grau, M. Chavarria, V. de Lorenzo, The interplay of the EIIA(Ntr)component of the nitrogen-related phosphotransferase system (PTS(Ntr)) of

2888 S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

Pseudomonas putida with pyruvate dehydrogenase, Biochim. Biophys. Acta10810 (2011) 995–1005.

[21] F. Velazquez, K. Pflüger, I. Cases, L.I. De Eugenio, V. de Lorenzo, Thephosphotransferase system formed by PtsP, PtsO, and PtsN proteins controls pro-duction of polyhydroxyalkanoates in Pseudomonas putida, J. Bacteriol. 189 (2007)4529–4533.

[22] I. Aranda-Olmedo, J.L. Ramos, S. Marques, Integration of signals through Crc andPtsN in catabolite repression of Pseudomonas putida TOL plasmid pWW0, Appl.Environ. Microbiol. 71 (2005) 4191–4198.

[23] I. Cases, J.A. Lopez, J.P. Albar, V. De Lorenzo, Evidence ofmultiple regulatory functions forthe PtsN (IIA(Ntr)) protein of Pseudomonas putida, J. Bacteriol. 183 (2001) 1032–1037.

[24] I. Cases, J. Perez-Martin, V. de Lorenzo, The IIANtr (PtsN) protein of Pseudomonasputida mediates the C source inhibition of the σ54-dependent Pu promoter of theTOL plasmid, J. Biol. Chem. 274 (1999) 15562–15568.

[25] I. Cases, F. Velazquez, V. de Lorenzo, Role of ptsO in carbon-mediated inhibition ofthe Pu promoter belonging to the pWWO Pseudomonas putida plasmid, J. Bacteriol.183 (2001) 5128–5133.

[26] T. del Castillo, J.L. Ramos, Simultaneous catabolite repression between glucose andtoluene metabolism in Pseudomonas putida is channeled through different signalingpathways, J. Bacteriol. 189 (2007) 6602–6610.

[27] F. Velazquez, I. di Bartolo, V. de Lorenzo, Genetic evidence that catabolites of theEntner–Doudoroff pathway signal C source repression of the sigma54 Pu promoterof Pseudomonas putida, J. Bacteriol. 186 (2004) 8267–8275.

[28] B. Zimmer, A. Hillmann, B. Görke, Requirements for the phosphorylation of theEscherichia coli EIIA(Ntr) protein in vivo, FEMS Microbiol. Lett. 286 (2008) 96–102.

[29] K. Pflüger, V. de Lorenzo, Evidence of in vivo cross talk between the nitrogen-relatedand fructose-related branches of the carbohydrate phosphotransferase system ofPseudomonas putida, J. Bacteriol. 190 (2008) 3374–3380.

[30] J. Reizer, A. Reizer, M.J. Merrick, G. Plunkett, D.J. Rose, M.H. Saier, Novelphosphotransferase-encoding genes revealed by analysis of the Escherichia coligenome: a chimeric gene encoding an Enzyme I homologue that possesses a puta-tive sensory transduction domain, Gene 181 (1996) 103–108.

[31] J. Reizer, A. Reizer, M.H. Saier, G.R. Jacobson, A proposed link between nitrogen andcarbon metabolism involving protein-phosphorylation in bacteria, Protein Sci. 1(1992) 722–726.

[32] K. Bettenbrock, S. Fischer, A. Kremling, K. Jahreis, T. Sauter, E.D. Gilles, A quantitativeapproach to catabolite repression in Escherichia coli, J. Biol. Chem. 281 (2006)2578–2584.

[33] T. Zeppenfeld, C. Larisch, J.W. Lengeler, K. Jahreis, Glucose transporter mutants ofEscherichia coli K-12 with changes in substrate recognition of IICB(Glc) and induc-tion behavior of the ptsG gene, J. Bacteriol. 182 (2000) 4443–4452.

[34] K.A. Datsenko, B.L. Wanner, One-step inactivation of chromosomal genes in Escherichiacoli K-12 using PCR products, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 6640–6645.

[35] J. Bjarnason, C.M. Southward, M.G. Surette, Genomic profiling of iron-responsivegenes in Salmonella enterica serovar typhimurium by high-throughput screeningof a random promoter library, J. Bacteriol. 185 (2003) 4973–4982.

[36] K. Bettenbrock, T. Sauter, K. Jahreis, A. Kremling, J.W. Lengeler, E.D. Gilles, Correla-tion between growth rates, EIIACrr phosphorylation, and intracellular cyclic AMPlevels in Escherichia coli K-12, J. Bacteriol. 189 (2007) 6891–6900.

[37] C.D. Herring, J.D. Glasner, F.R. Blattner, Gene replacementwithout selection: regulatedsuppression of amber mutations in Escherichia coli, Gene 311 (2003) 153–163.

[38] F.C. Neidhardt, P.L. Bloch, D.F. Smith, Culture medium for enterobacteria, J. Bacteriol.119 (1974) 736–747.

[39] S. Barik, Megaprimer PCR, in: B.-Y. Chen, H.W. Janes (Eds.), PCR Cloning Protocols,vol. 192, Humana Press, Totowa, 2002, pp. 189–196.

[40] J. Hellemans, G. Mortier, A. De Paepe, F. Speleman, J. Vandesompele, qBase relativequantification framework and software for management and automated analysisof real-time quantitative PCR data, Genome Biol. 8 (2007) R19.

[41] J. Vandesompele, K. De Preter, F. Pattyn, B. Poppe, N. Van Roy, A. De Paepe, F.Speleman, Accurate normalization of real-time quantitativeRT-PCRdata by geometricaveraging of multiple internal control genes, Genome Biol. 3 (2002) 12.

[42] E. Fischer, N. Zamboni, U. Sauer, High-throughput metabolic flux analysis based ongas chromatography–mass spectrometry derived 13C constraints, Anal. Biochem.325 (2004) 308–316.

[43] A. Nanchen, T. Fuhrer, U. Sauer, Determination of metabolic flux ratios from13C-experiments and gas chromatography–mass spectrometry data, in: W.Weckwerth (Ed.), Metabolomics: Methods and Protocols, vol. 358, Humana Press,Totowa, 2007, pp. 177–197.

[44] N. Zamboni, E. Fischer, U. Sauer, FiatFlux — a software for metabolic flux analysisfrom 13C-glucose experiments, BMC Bioinforma. 6 (2005) 209.

[45] N. Zamboni, S.M. Fendt, M. Ruhl, U. Sauer, C-13-based metabolic flux analysis, Nat.Protoc. 4 (2009) 878–892.

[46] H.B. Mann, D.R. Whitney, On a test of whether one of 2 random variables is stochas-tically larger than the other, Ann. Math. Stat. 18 (1947) 50–60.

[47] H. Takahashi, T. Inada, P. Postma, H. Aiba, CRP down-regulates adenylate cyclaseactivity by reducing the level of phosphorylated IIA(Glc), the glucose specificphosphotransferase protein, in Escherichia coli, Mol. Gen.Genet. 259 (1998) 317–326.

[48] M.M. Bradford, A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein-dye binding, Anal. Biochem.72 (1976) 248–254.

[49] J.R. Guest, I.T. Creaghan, Gene–protein relationships of alpha-keto acid dehydroge-nase complexes of Escherichia coli k12— isolation and characterization of lipoamidedehydrogenase mutants, J. Gen. Microbiol. 75 (1973) 197–210.

[50] R. Hengge-Aronis, D. Fischer, Identification and molecular analysis of glgS, a novelgrowth-phase-regulated and rpoS-dependent gene involved in glycogen-synthesisin Escherichia coli, Mol. Microbiol. 6 (1992) 1877–1886.

[51] B.D. Sak, A. Eisenstark, D. Touati, Exonuclease-Iii and the catalase hydroperoxidase-IIin Escherichia coli are both regulated by the Katf gene-product, Proc. Natl. Acad. Sci.U. S. A. 86 (1989) 3271–3275.

[52] A. Veit, T. Polen, V.F. Wendisch, Global gene expression analysis of glucose overflowmetabolism in Escherichia coli and reduction of aerobic acetate formation, Appl.Microbiol. Biotechnol. 74 (2007) 406–421.

[53] R. Noguez, D. Segura, S. Moreno, A. Hernandez, K. Juarez, G. Espin, Enzyme I-Ntr, NPrand IIA(Ntr) are involved in regulation of the poly-beta-hydroxybutyrate biosyn-thetic genes in Azotobacter vinelandii, J. Mol. Microbiol. Biotechnol. 15 (2008)244–254.

[54] T. Bahr, D. Lüttmann, W. Marz, B. Rak, B. Görke, Insight into bacterialphosphotransferase system-mediated signaling by interspecies transplanta-tion of a transcriptional regulator, J. Bacteriol. 193 (2011) 2013–2026.

[55] M. de Boer, C.P. Broekhuizen, P.W. Postma, Regulation of glycerol kinase byenzyme IIIglc of the phosphoenolpyruvate-carbohydrate phosphotransferasesystem, J. Bacteriol. 167 (1986) 393–395.

[56] T.P. Misko, W.J. Mitchell, N.D. Meadow, S. Roseman, Sugar-transport by the bacterialphosphotransferase system. Reconstitution of inducer exclusion in Salmonellatyphimurium membrane-vesicles, J. Biol. Chem. 262 (1987) 16261–16266.

[57] M.J. Novotny, W.L. Frederickson, E.B. Waygood, M.H. Saier, Allosteric regulation ofglycerol kinase by enzyme-IIIglc of the phosphotransferase system in Escherichiacoli and Salmonella typhimurium, J. Bacteriol. 162 (1985) 810–816.

[58] S.O. Nelson, J.K. Wright, P.W. Postma, The mechanism of inducer exclusion —direct Interaction between Purified IIIglc of the phosphoenolpyruvate — sugarphosphotransferase system and the lactose carrier of Escherichia coli, EMBO J.2 (1983) 715–720.

[59] T. Osumi, M.H. Saier, Regulation of lactose permease activity by the phosphoenol-pyruvate — sugar phosphotransferase system — evidence for direct binding of theglucose-specific enzyme III to the lactose permease, Proc. Natl. Acad. Sci. U. S. A.79 (1982) 1457–1461.

[60] N. Amin, A. Peterkofsky, A dual mechanism for regulating cAMP levels in Escherichiacoli, J. Biol. Chem. 270 (1995) 11803–11805.

[61] A. Peterkofsky, Y.J. Seok, N. Amin, R. Thapar, S.Y. Lee, R.E. Klevit, E.B. Waygood, J.W.Anderson, J. Gruschus, H. Huq, N. Gollop, The Escherichia coli adenylyl-cyclase com-plex — requirement of Pts proteins for stimulation by nucleotides, Biochemistry-Us34 (1995) 8950–8959.

[62] B.U. Feucht, M.H. Saier, Fine control of adenylate-cyclase by the phos-phoenolpyruvate — sugar phosphotransferase systems in Escherichia coliand Salmonella typhimurium, J. Bacteriol. 141 (1980) 603–610.

[63] B. Görke, J. Stülke, Carbon catabolite repression in bacteria: many ways to make themost out of nutrients, Nat. Rev. Microbiol. 6 (2008) 613–624.

[64] E. Gabor, A.K. Göhler, A. Kosfeld, A. Staab, A. Kremling, K. Jahreis, Thephosphoenolpyruvate-dependent glucose-phosphotransferase system fromEscherichia coli K-12 as the center of a network regulating carbohydrate flux inthe cell, Eur. J. Cell Biol. 90 (2011) 711–720.

[65] S.J. Park, J. McCabe, J. Turna, R.P. Gunsalus, Regulation of the citrate synthase (GltA)gene of Escherichia coli in response to anaerobiosis and carbon supply — role of theArcA gene-product, J. Bacteriol. 176 (1994) 5086–5092.

[66] S.J. Park, R.P. Gunsalus, Oxygen, iron, carbon, and superoxide control of the fumarasefumA and fumC genes of Escherichia coli — role of the ArcA, Fnr, and SoxRgene-products, J. Bacteriol. 177 (1995) 6255–6262.

[67] S.J. Park, P.A. Cotter, R.P. Gunsalus, Regulation of malate-dehydrogenase (mdh)gene-expression in Escherichia coli in response to oxygen, carbon, and heme avail-ability, J. Bacteriol. 177 (1995) 6652–6656.

[68] S.J. Park, C.P. Tseng, R.P. Gunsalus, Regulation of succinate-dehydrogenase(sdhCDAB) operon expression in Escherichia coli in response to carbon supply andanaerobiosis — role of ArcA and Fnr, Mol. Microbiol. 15 (1995) 473–482.

[69] S.J. Park, G. Chao, R.P. Gunsalus, Aerobic regulation of the sucABCD genes ofEscherichia coli, which encode alpha-ketoglutarate dehydrogenase and succinylcoenzyme A synthetase: roles of ArcA, Fnr, and the upstream sdhCDAB promoter,J. Bacteriol. 179 (1997) 4138–4142.

[70] G.L. Chao, J. Shen, C.P. Tseng, S.J. Park, R.P. Gunsalus, Aerobic regulation of isocitratedehydrogenase gene (icd) expression in Escherichia coli by the arcA and fnr geneproducts, J. Bacteriol. 179 (1997) 4299–4304.

[71] C.T. Gray, J.W. Wimpenny, M.R. Mossman, Regulation of metabolism in facultativebacteria. 2. effects of aerobiosis anaerobiosis and nutrition on the formation ofKrebs cycle enzymes in Escherichia coli, Biochim. Biophys. Acta 117 (1966) 33–41.

[72] M.Z. Liu, T. Durfee, J.E. Cabrera, K. Zhao, D.J. Jin, F.R. Blattner, Global transcriptionalprograms reveal a carbon source foraging strategy by Escherichia coli, J. Biol. Chem.280 (2005) 15921–15927.

[73] S.A. Bledig, T.M. Ramseier, M.H. Saier, FruR mediates catabolite activation of pyru-vate kinase (pykF) gene expression in Escherichia coli, J. Bacteriol. 178 (1996)280–283.

[74] T. Baba, T. Ara, M. Hasegawa, Y. Takai, Y. Okumura, M. Baba, K.A. Datsenko, M.Tomita, B.L. Wanner, H. Mori, Construction of Escherichia coli K-12 in-frame,single-gene knockout mutants: the Keio collection, Mol. Syst. Biol. 2 (2006) 1–11.

[75] C.R. Lee, B.M. Koo, S.H. Cho, Y.J. Kim,M.J. Yoon, A. Peterkofsky, Y.J. Seok, Requirementof the dephospho-form of enzyme IIA(Ntr) for derepression of Escherichia coli K-12ilvBN expression, Mol. Microbiol. 58 (2005) 334–344.

[76] B.M. Hogema, J.C. Arents, R. Bader, K. Eijkemans, H. Yoshida, H. Takahashi, H. Alba,P.W. Postma, Inducer exclusion in Escherichia coli by non-PTS substrates: the roleof the PEP to pyruvate ratio in determining the phosphorylation state of enzymeIIA(Glc), Mol. Microbiol. 30 (1998) 487–498.

[77] K. Kochanowski, B. Volkmer, L. Gerosa, B.R. Haverkorn van Rijsewijk, A. Schmidt, M.Heinemann, Functioning of a metabolic flux sensor in Escherichia coli, Proc. Natl.Acad. Sci. U. S. A. 110 (2013) 1130–1135.

2889S. Jahn et al. / Biochimica et Biophysica Acta 1833 (2013) 2879–2889

[78] C.D. Doucette, D.J. Schwab, N.S.Wingreen, J.D. Rabinowitz, alpha-ketoglutarate coor-dinates carbon and nitrogen utilization via enzyme I inhibition, Nat. Chem. Biol. 7(2011) 894–901.

[79] F. Mika, R. Hengge, A two-component phosphotransfer network involving ArcB,ArcA, and RssB coordinates synthesis and proteolysis of sigmaS (RpoS) in E. coli,Genes Dev. 19 (2005) 2770–2781.

[80] S. Iuchi, D.C. Cameron, E.C. Lin, A second global regulator gene (arcB) mediating re-pression of enzymes in aerobic pathways of Escherichia coli, J. Bacteriol. 171 (1989)868–873.

[81] S. Iuchi, E.C.C. Lin, Arca (Dye), a global regulatory gene in Escherichia colimediatingrepression of enzymes in aerobic pathways, Proc. Natl. Acad. Sci. U. S. A. 85 (1988)1888–1892.

[82] R.P. Gunsalus, S.J. Park, Aerobic-anaerobic gene regulation in Escherichiacoli: control by the ArcAB and Fnr regulons, Res. Microbiol. 145 (1994)437–450.

[83] E.J. Hansen, E. Juni, Two routes for synthesis of phosphoenolpyruvate fromC4-dicarboxylic acids in Escherichia coli, Biochem. Biophys. Res. Commun. 59(1974) 1204–1210.