Co-regulation of lipoamide dehydrogenase and 2-oxoglutaratedehydrogenase synthesis in Escherichia coli :

characterisation of an ArcA binding site in the lpd promoter

Louise Cunningham a, Dimitris Georgellis b, Je¡rey Green a, John R. Guest a;*a Department of Molecular Biology and Biotechnology, University of She¤eld, Western Bank, She¤eld S10 2TN, UK

b Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, MA 02115, USA

Received 2 October 1998; received in revised form 16 October 1998; accepted 19 October 1998

Abstract

The lipoamide dehydrogenase gene (lpdA) encoding the E3 subunits of both the pyruvate dehydrogenase and 2-oxoglutaratedehydrogenase complexes of Escherichia coli, is expressed from the upstream pdh and internal lpd promoters of the pdh operon(pdhR-aceEF-lpdA). Under aerobic conditions, the specific components of the 2-oxoglutarate dehydrogenase complex encodedby the sucAB genes in the sdhCDAB-sucABCD operon are expressed from the sdh promoter. The provision of lipoamidedehydrogenase subunits for assembly into the 2-oxoglutarate dehydrogenase complex could thus be controlled by co-regulationof the lpd promoter with the sdh promoter. Here, the transcription start point of the lpd promoter was defined by primerextension analysis, and an ArcA binding site, TGTTAACAAT, overlapping the lpd promoter and matching the consensus at8 out of 10 positions, was identified by in vitro footprint analysis. PdhR was not bound to the lpd promoter nor was ArcAbound specifically to the pdh promoter. These results support the view that co-regulation of the lpd and sdh promoters ismediated primarily by ArcA. z 1998 Federation of European Microbiological Societies. Published by Elsevier Science B.V.All rights reserved.

Keywords: Lipoamide dehydrogenase; Pyruvate dehydrogenase complex; 2-Oxoglutarate dehydrogenase complex; ArcA; Transcription

regulation; Escherichia coli

1. Introduction

The pyruvate dehydrogenase (PDH) complex ofEscherichia coli contains multiple copies of three sub-units: pyruvate dehydrogenase (E1p); lipoate acetyl-transferase (E2p); and lipoamide dehydrogenase(E3). The corresponding genes are located at 2.7min in the pdh operon (pdhR-aceE-aceF-lpdA ; Fig. 1)

which contains two major promoters: Ppdh, generat-ing a 7.4-kb pdhR-lpdA readthrough transcript; andPlpd, which generates an independent 1.7 kb lpdAtranscript [1,2]. The pdh promoter is negatively au-toregulated by PdhR, which binds to a speci¢c oper-ator sequence in the absence of pyruvate [1]. Thegenes encoding the speci¢c dehydrogenase (E1o,sucA) and succinyltransferase (E2o, sucB) subunitsof the analogous 2-oxoglutarate dehydrogenase(ODH) complex are located at 16.3 min in the sdh-

0378-1097 / 98 / $19.00 ß 1998 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 3 7 8 - 1 0 9 7 ( 9 8 ) 0 0 5 0 6 - 0

FEMSLE 8505 3-12-98

* Corresponding author.

FEMS Microbiology Letters 169 (1998) 403^408

suc operon (sdhCDAB-sucABCD) which also encodessuccinate dehydrogenase (SDH; sdhCDAB) and suc-cinyl-CoA synthetase (sucCD). These genes are ex-pressed primarily from the sdh promoter, but there isan internal suc promoter which seems to be respon-sible for the anaerobic and stationary phase expres-sion of the sucABCD genes [3,4]. The single lpdAgene is responsible for providing E3 subunits forassembly into both the PDH and ODH complexes.These complexes are di¡erentially regulated and themechanism which co-ordinates the synthesis of theE3 subunits with the individual requirements ofeach complex emerged from studying transcript syn-thesis under di¡erent growth conditions [2]. Thus itappeared that for synthesis of the PDH complex, thelpdA gene is expressed together with the aceEF genesfrom Ppdh as the distal gene of the pdh operon,whereas for synthesis of the ODH complex it is ex-pressed from the independent lpd promoter which inturn is co-regulated with the sucA and sucB genes,now known to be expressed from Psdh [3,4].

Enzymological studies with an arcA mutant origi-nally showed that the PDH and ODH complexes(and SDH) are anaerobically repressed by ArcA[5]. In the case of the PDH complex, this couldmean that either one or both of the two relevantpromoters (Ppdh and Plpd) might be regulated byArcA. Subsequent studies with lacZ fusions indi-cated that Ppdh is una¡ected by arcA mutationwhereas Plpd activity is anaerobically derepressedby arcA mutation [1]. The latter promoter (Plpd)has also been shown to be activated by Fis and re-pressed by glucose via an ill-de¢ned CRP-independ-ent mechanism [4]. In common with Plpd, the sdhpromoter has been shown to be anaerobically re-pressed by ArcA, aerobically activated by Fis, andsubject to CRP-independent repression by glucose[6], although there are also reports that Psdh is re-pressed by Fis [7] and subject to CRP-dependentglucose repression [8]. Thus, in the absence of glu-cose, ArcA and Fis are the most plausible candidatesfor mediating the co-regulation of lipoamide dehy-drogenase and 2-oxoglutarate dehydrogenase synthe-sis via the respective lpd and sdh promoters. This hasbeen investigated here by mapping the start point oflpd transcription, and by in vitro studies on the bind-ing of ArcA, Fis and PdhR to the lpd promoter, andof ArcA to the pdh promoter.

2. Materials and methods

2.1. Bacterial strains and plasmids

E. coli DH5K was the routine transformation hostand W3110 (prototroph) was used as the RNAsource. Promoter-containing DNA was obtainedfrom two plasmids, pGS698 for Ppdh [1] andpGS704 for Plpd [9] (Fig. 1). Two plasmids,pGS1146 and pGS1147, were constructed by sub-cloning the 0.55-kb RsaI fragment containing thelpd promoter from pGS704 into the SmaI site ofpUC119 and de¢ning the orientation of the insertby locating the asymmetric BglI site (Fig. 1).

2.2. Microbiological methods, DNA manipulation,protein puri¢cation and DNase I footprinting

Strains were cultured aerobically in L-broth withampicillin (100 Wg ml31) where required, and DNAwas isolated and manipulated by standard methods[10]. His6-ArcA protein was puri¢ed and phosphoryl-ated as described previously [11] and Fis protein wasprepared from an ampli¢ed source, RJ2465 [12]. The

FEMSLE 8505 3-12-98

Fig. 1. Transcriptional organisation of the pdh operon (pdhR-aceEF-lpdA). The pdh and lpd promoters are shown and the cor-responding major transcripts are denoted by horizontal arrows.Potential stem-loop structures, including an intergenic repeat unit(IRU) or enterobacterial repetitive intergenic consensus (ERIC)immediately upstream of the lpdA coding region, are indicated inthe intergenic regions. The plasmids used as DNA sources andrelevant (but not necessarily unique) restriction sites are alsoshown: B, BamHI; Bg, BglI; D, DraI; E, EcoRI; H, HindIII; P,PstI; R, RsaI; and X, XhoII; subscript v refers to £anking vec-tor sites.

L. Cunningham et al. / FEMS Microbiology Letters 169 (1998) 403^408404

PdhR protein was isolated from BL21DE3(pGS680)by a modi¢cation of the published method [13]. Cul-tures in L-broth were induced with IPTG (0.1 mgml31) for 90 min after growth to OD600 = 0.6.PdhR was ¢rst enriched from cell-free extracts byheparin-agarose chromatography, applying the sam-ples in bu¡er A (20 mM Tris-HCl, 1 mM EDTA,1 mM MgCl2, 1 mM sodium azide, 1 mM DTT,0.1 mM PMSF, 10% glycerol, ¢nal pH 7.3) and elut-ing with an ammonium sulfate gradient (0^1 M inbu¡er A). The ammonium sulfate concentration offractions containing PdhR were increased from0.45 to 2.0 M for Phenyl Superose reverse-phasechromatography with an ammonium sulfate gradient(2-0 M in bu¡er A). The PdhR protein was elutedat 1.2 M and diluted for use without removing thesalt.

For DNase I footprinting [13], 2-pmol samples ofDNA (end-labelled with appropriate [K-32P]dNTP)were incubated for 10 min at 30³C with 0.5^4.0WM phosphorylated ArcA, 4 Wl of 5U bandshiftbu¡er (0.5 M Tris-Cl pH 7.4, 0.5 M KCl, 50 mMMgCl2, 50% glycerol, 10 mM DTT; [11]), 10 WgBSA and 0.2 Wg sheared calf thymus DNA (20 Wl¢nal volume), prior to adding MgCl2 (8 mM, ¢nalconcentration) and digesting with 3 Wl DNase I (10U/ml; Boehringer Mannheim) for 1 min at 30³C.After ethanol precipitation, with 20 Wg glycogen ascarrier, samples were washed with 70% ethanol,dried under vacuum and fractionated in a 5% acryl-amide 7 M urea sequencing gel, and then analysedby autoradiography.

2.3. RNA extraction and primer extension analysis

The hot acid phenol procedure [14] was used toextract RNA from exponential phase cultures(OD600 = 0.6) that had been rapidly cooled to 4³Cin liquid N2 [15]. The method used for primer exten-sion analysis [16] was modi¢ed to allow continuousincorporation of [K-32P]dCTP [15]. Samples of totalRNA (100 Wg) from cultures of W3110 harvested atOD600 = 0.5^0.6, were used with 10 pmol primer andfractionated after processing by electrophoresis in6% acrylamide/7 M urea gels alongside sequence lad-ders derived from the corresponding DNA and prim-er. The oligonucleotide primers for analysing lpdtranscripts were: S560 (5892^5868) and S561

(6017^5994); co-ordinates based on GenBankV01498.

3. Results and discussion

3.1. Identi¢cation of the start points of lpd mRNA byprimer extension analysis

Primer extension analysis with RNA from mid-ex-ponential phase cultures of E. coli strain W3110 wasused to obtain a precise identi¢cation of the startpoint of lpdA transcription (Fig. 2a). Transcriptionwas shown to start primarily at the ¢rst of two ad-jacent G nucleotides (co-ordinates 5808 and 5809;GenBank V01498) just 7 bp downstream from theputative 310 motif of the lpd promoter (Fig. 2ai;Fig. 3). The same start point was identi¢ed with asecond oligonucleotide that primes from a di¡erentsite (not shown). This start point is more optimallyplaced than the T and G nucleotides located 3 and 5bp further downstream, that were tentatively as-signed in previous studies using a less precise S1mapping procedure [2]. Four larger extension prod-ucts starting at four consecutive nucleotides, TCTC(co-ordinates 5718^5721; GenBank V01498), werealso detected in the same primer extension reactions(Fig. 2aii). They correspond to mRNA moleculesthat start 8^11 bp downstream of a potential stem-loop, located immediately distal to the aceF codingregion, and continue through the lpd promoter andthe lpdA coding region to the lpd terminator (Fig.2aii ; Fig. 3). These transcripts probably arise by en-donucleolytic cleavage of the major pdhR-lpdA read-through transcript (Fig. 1). Evidence for such proc-essing in the aceF-lpdA intergenic region wasobserved previously [2] when mRNA molecules ter-minating at three sites lying within or close to thestem-loop (marked by ¢lled circles in Fig. 3) weredetected by S1 mapping. At the time, these mRNAmolecules were thought to be formed by transcrip-tion termination immediately downstream of theaceF gene. However, no major transcripts terminat-ing in the aceF-lpdA intergenic region can be de-tected by Northern blotting [1]. The possibility thatthe longer lpdA transcripts are generated by a secondlpd promoter, located within the sequence encodingthe stem-loop, seems unlikely because this sequence

FEMSLE 8505 3-12-98

L. Cunningham et al. / FEMS Microbiology Letters 169 (1998) 403^408 405

lacks typical promoter elements. Nevertheless, theexistence of such a promoter has not been ruled out.

3.2. Studies on the binding of ArcA, Fis and PdhR atthe lpd promoter

DNase I footprinting with phosphorylated ArcAwas used with two end-labelled 573 bp EcoRI-Bam-HI fragments from pGS1146 and pGS1147 (Fig. 1)to detect potential ArcA binding sites associatedwith the lpd promoter (Fig. 2b). A single site ofprotection was detected with both strands (Fig.2b), the protection being typically stronger with thecoding strand (Kd = 2.0 WM) than the non-codingstrand (Kd = 2.7 WM). The footprints of the codingstrand identify a hyper-sensitive site just upstream ofthe ArcA binding site (Fig. 2b), suggesting that theconformation of the DNA is perturbed when the

regulator is bound. The protected region overlapsthe 310 and 335 hexamers of Plpd, consistent withArcA being an anaerobic repressor of lpdA gene ex-pression [4]. Furthermore, the protected region con-tains a sequence that matches the ArcA binding-siteconsensus, WGTTAATTAW [11], at 8 out of 10positions (Fig. 3). The presence of a single ArcAsite at the lpd promoter contrasts markedly withthe presence of four ArcA binding sites associatedwith the sdh promoter [17]. It is also interestingthat the ArcA-site in the lpd promoter is located inan 18-bp palindromic sequence, AAAATTGTTAA-CAATTTT (Fig. 3), which is unique in the E. coligenome. There are eight closely related sequences,each having three mismatches, and two such sequen-ces located upstream of the dnaK and yaaI genesexhibit hyphenated dyad symmetry of the type:AAAATTGnnnnCAATTTT.

FEMSLE 8505 3-12-98

Fig. 2. Location of the transcription start site and the ArcA binding site of the lpd promoter. (a) Primer extension analysis with RNA ex-tracted from E. coli W3110 (prototroph) and one of the lpd primers, S560. The major products (marked by asterisks) observed in down-stream (i) and upstream (ii) regions of the same autoradiograph are aligned with a sequence ladder generated from pGS704 DNA withthe same primer. The positions of the start sites deduced previously by S1 mapping [2] are shown in bold and underlined. The same prod-ucts were observed with a di¡erent primer, S561. (b) DNase I footprint analyses showing the segments of the lpd promoter region thatare protected by phosphorylated ArcA. The coding and non-coding strands of the 573-bp BamHI-EcoRI fragments of pGS1146 andpGS1147 (respectively) were labelled with [K-32P]dATP at their EcoRI sites, and incubated with phosphorylated ArcA prior to digestionwith DNase I and electrophoretic fractionation. In each case, the concentrations of phosphorylated ArcA (WM) were (left to right) : noprotein; 0.5; 1.0; and 2.0). The transcription start-sites (+1), the 310 and 335 hexamers, the nucleotide co-ordinates (GenBank V01498),and the ArcA-protected regions and a hyper-sensitive site (3), are shown.

L. Cunningham et al. / FEMS Microbiology Letters 169 (1998) 403^408406

Puri¢ed Fis protein (v 1.5 nM) retarded the same573 bp lpd promoter fragment in gel-retardation as-says. However, no protection could be detected byDNase I footprint analysis using up to 30 nM Fis(data not shown). This suggests that the observedretardation is probably due to non-speci¢c Fis bind-ing and it was concluded that the e¡ects of Fis onlpdA expression may be indirect.

Evidence that the pyruvate-sensing repressor(PdhR) might regulate Plpd came from the 40% in-crease in L-galactosidase activity observed when anlpdA-lacZ fusion lacking the proximal segment of thepdh operon, is expressed in a pdhR deletion strain [1].However, puri¢ed samples of PdhR protein, whichwere shown to bind speci¢cally to Ppdh DNA in gelretardation and footprinting studies, had no e¡ect onthe lpd promoter fragment in either gel retardationassays or footprint analysis under comparable condi-tions (data not shown). This is consistent with theabsence of a potential PdhR binding site in the lpdpromoter region. These observations suggest that theincrease in independent lpdA-lacZ expression in thePdhR-de¢cient strain is an indirect e¡ect and it wastherefore concluded that PdhR controls the synthesisof the PDH complex solely by regulating Ppdh activity.

3.3. Attempts to demonstrate ArcA-binding at thePdhR promoter

Studies with pdhR-lacZ fusions indicated that Ppdh

is not regulated by ArcA [1] even though a 1.7-foldanaerobic derepression of pyruvate dehydrogenasesynthesis was observed in an arcA mutant [5]. The541 bp BamHI^HindIII fragment of pGS680 con-taining the pdh promoter (Fig. 1) was accordinglylabelled with [K-32P]dGTP and used in gel retarda-tion assays and footprint analysis. PhosphorylatedArcA retarded the Ppdh fragment, but only in directproportion to the protein concentration, indicatingthat the binding is non-speci¢c. There are three se-quences within þ 50 bp of the pdh transcription startpoint that match the ArcA-site consensus at 7 or 8out of 10 positions. However, footprinting studiesshowed that the pdh promoter region is not protectedby phosphorylated ArcA (data not shown). There-fore, based on this in vitro evidence, it was con-cluded that Ppdh is not regulated by ArcA. Presum-ably the increase in PDH activity observed in thearcA mutant is a secondary consequence of thearcA mutation.

FEMSLE 8505 3-12-98

Fig. 3. Diagrammatic representation of the aceF-lpdA intergenic region containing the independent lpd promoter. The sequence (co-ordi-nates 5675^6005; GenBank V01498) is bounded by the aceF stop and the lpdA start codons (underlined) shows the transcription start site(+1, C) and the 310 and 335 hexamers (bold face) of the lpd promoter. The regions of ArcA protection are overlined (coding stand) orunderlined (non-coding strand) and an ArcA binding-site consensus [11] is shown above the sequence; W = T/A. Also shown are the posi-tions of an 18-bp palindrome and a 125-bp intergenic repeat unit (IRU) or enterobacterial repetitive intergenic consensus (ERIC) by con-verging arrows, a potential stem-loop structure (mRNA processing site or aceF transcription terminator) with ¢lled circles to denote the3P-ends of aceF transcripts de¢ned previously by S1 mapping [2]. The vertical arrows mark the 5P-ends of lpdA transcripts generated byendonucleolytic mRNA processing of aceF-lpdA readthrough transcripts or the products of a second lpd promoter, detected by primer ex-tension analysis.

L. Cunningham et al. / FEMS Microbiology Letters 169 (1998) 403^408 407

3.4. Conclusions

Previous studies with an lpdA-lacZ fusion lackingthe upstream genes have shown that lpd gene expres-sion responds mainly to ArcA-mediated anaerobicrepression, Fis-mediated activation under aerobicconditions, and CRP-independent repression by glu-cose [4]. In the present in vitro studies, no speci¢cbinding site was detected for Fis, suggesting that itse¡ect is indirect. However, a single binding site forphosphorylated ArcA, overlapping the 335 and 310hexamers, was identi¢ed in the lpd promoter. Thisstrongly suggests that in response to anaerobiosis,the formation and subsequent binding of phosphory-lated ArcA simply blocks RNA polymerase bindingand hence represses independent transcription of thelpdA gene. It was thus concluded that the co-expres-sion of lipoamide dehydrogenase and 2-oxoglutaratedehydrogenase involves co-regulation of the lpd andsdh promoters, mediated primarily by ArcA and anill-de¢ned CRP-independent mechanism of glucoserepression. PdhR, the pyruvate-sensitive repressorthat crucially controls the synthesis of the PDH com-plex via Ppdh, also failed to bind to Plpd, indicatingthat although PdhR regulates the synthesis of lipoa-mide dehydrogenase via Ppdh, it does not regulatePlpd. The in vitro studies also con¢rmed that Ppdh

activity is not regulated by ArcA.

Acknowledgments

We thank Dr. Ed Lin for helpful advice. The workwas supported by a project grant from The Well-come Trust (J.R.G.), a research fellowship (B-PD11474-301) from the Swedish National Science Re-search Council (D.G.), and a BBSRC Advanced Fel-lowship (J.G.).

References

[1] Quail, M.A., Haydon, D.J. and Guest, J.R. (1994) The pdhR-aceEF-lpd operon of Escherichia coli expresses the pyruvatedehydrogenase complex. Mol. Microbiol. 12, 95^104.

[2] Spencer, M.E. and Guest, J.R. (1985) Transcription analysisof the sucAB, aceEF and lpd genes of Escherichia coli. Mol.Gen. Genet. 200, 145^154.

[3] Park, S.-J., Chao, G. and Gunsalus, R.P. (1997) Aerobic reg-ulation of the sucABCD genes of Escherichia coli, which en-code K-ketoglutarate dehydrogenase and succinyl coenzyme Asynthetase: roles of ArcA, Fnr, and the upstream sdhCDABpromoter. J. Bacteriol. 179, 4138^4142.

[4] Cunningham, L. and Guest, J.R. (1998) Transcription andtranscript processing in the sdhCDAB-sucABCD operon ofEscherichia coli. Microbiology 144, 2113^2123.

[5] Iuchi, S. and Lin, E.C.C. (1988) arcA (dye), a global regula-tory gene in Escherichia coli mediating repression of enzymesin aerobic pathways. Proc. Natl. Acad. Sci. USA 85, 1888^1892.

[6] Park, S.-J., Tseng, C.-P. and Gunsalus, R.P. (1995) Regula-tion of succinate dehydrogenase (sdhCDAB) operon expres-sion in Escherichia coli in response to carbon supply andanaerobiosis : role of ArcA and Fnr. Mol. Microbiol. 15,473^482.

[7] Xu, J. and Johnson, R.C. (1995) Identi¢cation of genes neg-atively regulated by Fis: Fis and RpoS comodulate growth-phase-dependent gene expression in Escherichia coli. J. Bacter-iol. 177, 938^947.

[8] Wilde, R.J. and Guest, J.R. (1985) Transcript analysis of thecitrate synthase and succinate dehydrogenase genes of Esche-richia coli K12. J. Gen. Microbiol. 132, 3239^3251.

[9] Quail, M.A., Dempsey, C.E. and Guest, J.R. (1994) Identi¢-cation of a fatty acyl responsive regulator (FarR) in Esche-richia coli. FEBS Lett. 356, 183^187.

[10] Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) MolecularCloning: A Laboratory Manual, 2nd edn. Cold Spring HarborLaboratory Press, New York.

[11] Lynch, A.S. and Lin, E.C.C. (1996) Transcriptional controlmediated by the ArcA two-component response regulator pro-tein of Escherichia coli : characterisation of DNA binding attarget promoters. J. Bacteriol. 178, 6238^6249.

[12] Osuna, R., Finkel, S.E. and Johnson, R.C. (1991) Identi¢ca-tion of two functional regions of Fis: the N-terminus us re-quired to promote Hin-mediated DNA inversion but not ex-cision. EMBO J. 10, 1593^1603.

[13] Quail, M.A. and Guest, J.R. (1995) Puri¢cation, characteriza-tion and mode of action of PdhR: the transcriptional repress-or of the pdhR-aceEF-lpd operon of Escherichia coli. Mol.Microbiol. 15, 519^529.

[14] Aiba, H., Adhya, S. and de Crombrugghe, B. (1981) Evidencefor two functional gal promoters in intact Escherichia colicells. J. Biol. Chem. 256, 11905^11910.

[15] Cunningham, L., Gruer, M.J. and Guest, J.R. (1997) Tran-scriptional regulation of the aconitase genes (acnA and acnB)of Escherichia coli. Microbiology 143, 3795^3805.

[16] Gerischer, U. and Duërre, P. (1992) mRNA analysis of the adcgene region of Clostridium acetobutylicum during the shift tosolventogenesis. J. Bacteriol. 174, 426^433.

[17] Shen, J. and Gunsalus, R.P. (1995) Role of multiple ArcArecognition sites in anaerobic regulation of succinate dehydro-genase (sdhCDAB) gene expression in Escherichia coli. Mol.Microbiol. 15, 519^529.

FEMSLE 8505 3-12-98

L. Cunningham et al. / FEMS Microbiology Letters 169 (1998) 403^408408