JOURNAL OF BACTERIOLOGY, Sept. 1992, p. 5617-5623 Vol. 174, No. 170021-9193/92/175617-07$02.00/0Copyright © 1992, American Society for Microbiology

Purification and Phosphorylation of the Arc RegulatoryComponents of Eschenichia coli

S. IUCHI AND E. C. C. LINDepartment ofMicrobiology and Molecular Genetics, Harvard Medical School,

200 Longwood Avenue, Boston, Massachusetts 02115

Received 30 March 1992/Accepted 9 June 1992

In Escherichia coli, a two-component signal transduction system, consisting of the transmembrane sensorprotein ArcB and its cognate cytoplasmic regulatory protein ArcA, controls the expression of genes encodingenzymes involved in aerobic respiration. ArcB belongs to a subclass of sensors that have not only a conservedhistidine-containing transmitter domain but also a conserved aspartate-containing receiver domain of theregulator family. 'ArcB (a genetically truncated ArcB missing the two transmembrane segments on theN-terminal end) and ArcA were purified from overproducing cells. Autophosphorylation of 'ArcB was revealedwhen the protein was incubated with [_y-32P]ATP but not with [a-32PJATP or [.y-32PJGTP. When ArcA wasincubated in the presence of 'ArcB and [_y-32PJATP, ArcA acquired radioactivity at the expense of thephosphorylated protein 'ArcB-32P. When a limited amount of 'ArcB was incubated with excess ArcA and[_y_32PJATP, ArcA-32P increased linearl with time. Under such conditions, for a given time period the amountof ArcA phosphorylated was proportional to the concentration of 'ArcB. Thus, 'ArcB acted as a kinase forArcA. Chemical stabilities of the phosphorylated proteins suggested that 'ArcB-32P contained both a histidylphosphate and an aspartyl phosphate(s) and that ArcA-32P contained only an aspartyl phosphate(s).

In Escherichia coli, products of the arcA (min 0) and arcB(min 69.5) genes pleiotropically control the expression ofnumerous target operons, exerting both positive and nega-tive effects. For instance, expression of the structural genesencoding several dehydrogenases of the flavoprotein class;the cytochrome o complex; and members of the tricarboxy-lic acid cycle, the glyoxylate shunt, and the pathways forfatty acid degradation are anaerobically repressed (15, 17).In contrast, expression of the cydAB operon encoding the02-scavenging cytochrome d is activated under anoxic con-ditions (6, 16). As a network, these enzymes are responsiblefor the efficient generation of metabolic energy by aerobicrespiration. ArcA (or SfrA, for sex factor regulation) is alsorequired for expression of the tra (DNA transfer) genes ofthe F plasmid. In this function, CpxA, instead of ArcB,might act in conjunction with ArcA (18, 30, 32, 33, 38).ArcA and ArcB belong to a family of homologous two-

component regulatory systems in eubacteria. Typically,such a system comprises (i) a sensor protein with a con-served transmitter domain containing a His and (ii) a regu-lator protein with a receiver domain containing a cluster ofAsp residues. In almost every case, the sensor is a trans-membrane protein, whereas the regulator is a cytoplasmicprotein (4, 20, 23, 29, 30, 35). It has been shown for severalsystems that upon stimulation, the sensor undergoes auto-phosphorylation at a conserved His residue. The phosphorylgroup is then transferred to a conserved Asp residue(s) of thecognate regulator, whereupon it becomes functionally active(1, 2, 5, 7, 11-13, 21, 22, 24, 26-28, 31, 35, 39). However,phosphorylation of Arc components has not yet been dem-onstrated. The sensor ArcB, interestingly, possesses notonly a transmitter domain in the N-terminal region (includinga conserved His-292) but also a receiver domain in theC-terminal region (including conserved Asp-533 and Asp-576) (Fig. 1) (20). Such an elaborate sensor is not excep-tional. Several other sensors are known to possess both atransmitter and a receiver domain (3, 34-36). In the case ofthe sensor VirA of Agrobacterium tumefaciens, a His resi-

due has been shown to undergo autophosphorylation (21),and in the case of the sensor FrzE of Myxococcus xanthus,an Asp residue(s) has been shown to undergo autophospho-rylation (26). In contrast, ArcA is a typical regulator protein,with only the receiver domain containing conserved Aspresidues at positions 11 and 54 as potential sites fortransphosphorylation by phosphorylated ArcB (ArcB-P) (4,20).

In this study, we constructed plasmids for producing andpurifying the two proteins 'ArcB (an ArcB lacking the twotransmembrane portions in the N-terminal region) and ArcAfor the purpose of testing (i) whether ArcB undergoesautophosphorylation at both His and Asp residues and (ii)whether an Asp residue of ArcA can receive a phosphorylgroup from ArcB-P.

MATERUILS AND METHODS

Bacteria, phages, and plasmids. All strains used (XL1-Blue[Stratagene], K38, and ECL594) were derivatives of E. coliK-12. P1 vir phage was used for transductions. To placearcBl in strain K38 with a plasmid (pGP1-2) encoding the T7polymerase gene (37), the mutant allele in strain ECL594 (15)was cotransduced with a closely linked TnlO (85%). Tetrtransductants were selected on LB agar containing tetracy-cline and purified on the same type of agar. Coinheritance ofarcBl was scored for sensitivity to the redox dye toluidineblue o (20). Plasmid pT7-7 was provided by S. Tabor.Growth conditions. For routine cultures, LB medium (1%

tryptone, 0.5% yeast extract, 0.5% NaCl) was used. Toselect Tetr transductants, tetracycline was added at 10 ,ug/mlin LB agar. Ampicillin (100 ,ug/ml) and kanamycin (75 ,ug/ml)were added to LB medium to prevent loss of plasmid vectorsbearing a drug resistance marker. To test for dye sensitivity,toluidine blue o was added at 0.2 mg/ml in tryptone agar (20).DNA manipulation and transformation. Plasmid DNA was

prepared on a small scale by the boiling method (25).Bacterial cells were transformed by the CaCl2 method (25).

5617

on July 8, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

5618 IUCHI AND LIN

100 AA

ArcB

'ArcB

(%"4

s 000I1-- I

a

(%"4N ~ nil)

CN Le) U')

0 0

Lo

ArcAFIG. 1. Schematic diagrams of the ArcB, 'ArcB, and ArcA

proteins. Dark boxes indicate putative transmembrane portions.Hatched boxes indicate the conserved transmitter domain; openboxes indicate the conserved receiver domain; numbers over thediagrams indicate the positions of the histidyl (H) and aspartyl (D)residues.

Estimation of protein concentration. The Coomassie pro-tein assay reagent (Pierce) was used to estimate proteinconcentrations, with bovine serum albumin as a standard.

Purification of 'ArcB. Strain K38/pGP1-2 was used as thehost strain to hyperproduce 'ArcB. The cells were trans-formed with pTB5 (Fig. 2) and grown undisturbed overnightat 30°C in 200 ml of LB medium containing kanamycin andampicillin. The culture was then heated in a 42°C water bathfor 5 min and shaken for 90 min at 37°C. After being chilled,the cells were harvested by centrifugation. The pellet wasresuspended in 50 ml of 20 mM 3-(N-morpholino)propanesulfonate (MOPS) buffer (pH 7.6) containing 1 mM EDTAand centrifuged. The final pellet was resuspended in 1 ml ofthe same buffer for sonic disruption, and the sonic extract

K

pBB22 L

PTB5

pBA2

p vI

arcB

T7 SD IP V H

5. 3'arcB

H V KT7I _ 3s' 3.

arcAFIG. 2. Schematic diagrams of plasmids bearing the arcB, 'arcB,

or arcA gene. Solid lines indicate inserts, and thickened portionsindicate the coding regions. Dotted lines indicate part of the vectorDNA. Abbreviations: K, KpnI; P, PstI; V, EcoRV; H, HindIII;17,promoter; SD, Shine-Dalgarno sequence.

was centrifuged for 60 min at 100,000 x g. The membranepellet fraction, containing an amount of the protein approx-imately equal to that in the supernatant fraction, was used assource of 'ArcB for further purification, since there wasrelatively little contamination by cytoplasmic proteins. Aftersuspension of the membrane fraction in 1 ml of MOPS-EDTA buffer, the preparation was subjected to sodiumdodecyl sulfate (SDS)-polyacrylamide gel electrophoresis(PAGE) (5% acrylamide gel) fractionation as described pre-viously (8). After completion of the electrophoresis, theprotein in the gel was revealed by a cold solution of 0.25 MKCI and 1 mM dithiothreitol (DTT). The visible band con-taining the protein was cut out and extracted with Electro-Eluter (Bio-Rad). The extract was dried with acetone. Thematerial was dissolved in 50 mM Tris buffer (pH 8.0)containing 1 mM EDTA, 5 mM DTT, and 6 M guanidine, andthe proteins were renatured by overnight dialysis against 200volumes of 20 mM N-(2-hydroxyethyl)piperazine-N'-(2-ethanesulfonate) (HEPES) buffer (pH 7.5) containing 50 mMKCl, 5 mM DTT, and 50% glycerol (21). The purified proteinwas kept at -20°C until used.

Purification of ArcA. Strain K38/pGP1-2 was also used asthe host strain to hyperproduce ArcA. The cells weretransformed with plasmid pBA2 (Fig. 2). The above-de-scribed procedure for preparing 'ArcB was followed throughthe step of 100,000 x g centrifugation. This time, the solublefraction which contained almost all of the ArcA protein wasrecovered instead of the pellet and diluted sixfold withMOPS-EDTA buffer. Proteins in the solution were precip-itated by ammonium sulfate at 60% saturation and 0°C. Theprecipitate was dissolved in 1 ml of MOPS-EDTA buffer anddialyzed overnight against 1 liter of the same buffer. Thedialysate (about 2 ml) was applied to a column of DE52 anionexchanger (14-ml bed volume) equilibrated with MOPS-EDTA buffer. The proteins were eluted with 30 ml ofMOPS-EDTA buffer containing 140 mM NaCl. The eluatewas concentrated with Centriprep-30 (Amicon) to 0.4 ml andapplied to an SDS-polyacrylamide (10% acrylamide) gel.This was the most effective step in the purification proce-dure. The purified ArcA was recovered and renatured asdescribed for 'ArcB. To remove some remaining contami-nating proteins, the ArcA preparation was applied to ahydroxyapatite column (2-ml bed volume) equilibrated with10 mM potassium phosphate buffer (pH 6.8). The proteinswere serially eluted with 2 ml of potassium phosphate bufferin which the concentration of NaCl was increased by 10 mMat each step from 50 to 160mM. Fractions of 110, 120, and130 mM NaCl, containing most of the total ArcA, werecombined. (Most of the other proteins were eluted in thefractions of 50 to 60 mM NaCl.) The combined ArcA-containing fraction was concentrated with Centricon 10(Amicon) and dissolved in 0.4 ml of the HEPES-KCl-DTT-20% glycerol buffer (pH 7.5). The purified ArcA was kept at-20°C until used.Phosphorylation of proteins. Unless otherwise specified,

protein phosphorylations were carried out at 25°C in 10-,ureaction mixtures containing 0.1 mM [,y-32P]ATP (500 cpm/pmol), 33 mM HEPES (pH 7.5), 70 mM KCI, 10mM MgCl2,2 mM DTT, and 0.1 mM EDTA. The reaction was initiatedby addition of the radioactive substrate and stopped byaddition of equal volume of twice-concentrated loadingbuffer (50 mM Tris-HCl [pH 6.8], 4% SDS, 10% glycerol, 5%2-mercaptoethanol, 0.003% bromophenol blue) for electro-phoresis. Terminated reaction mixtures were promptlyheated for 3 min at 55°C (13). In time course experimentscarried out in larger reaction volumes, samples mixed with

J. BACTERIOL.

on July 8, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHORYLATION OF Arc REGULATORY PROTEINS 5619

the loading buffer were kept on ice until the last portion wastaken. The denatured 'ArcB and/or ArcA proteins wereseparated by SDS-PAGE (10% polyacrylamide gel, pH 8.8)in a MINI-2D gel apparatus (Bio-Rad). After completion ofthe electrophoresis, the gel was stained by Coomassie BlueR for 20 min, destained for 20 min twice, and dried to revealthe protein bands. Labeling of proteins in the gel wasdetermined qualitatively by autoradiography with X-OmatAR diagnostic film (Kodak). A Phosphorimager (MolecularDynamics) was used for quantitative determination of radio-activity, and the intensity was expressed in units of intensityper minute (exposure time varied from 15 to 40 h).

Stability of the phosphate links of 'ArcB and ArcA. Tocompare the chemical stability of the phosphate links to theproteins, we terminated the phosphorylation reaction byadding the loading buffer but omitting the 55°C heat treat-ment. The 'ArcB protein (30 pmol) was phosphorylated for2.5 min in 30 ,ul of the reaction mixture. Samples of 10 ,ul(each originally containing 5 pmol of 'ArcB) were pipettedinto microcentrifuge tubes containing 2 ,ul of 1 N HCl, 0.15N HCI, water, or 3 N NaOH, giving a final pH of 0.9, 2.7,7.2, or 12.5, respectively (5). The mixtures were then incu-bated for 60 min in a water bath at 43°C. Another sample,without pH adjustment, was kept on ice as a control. Thesamples were separated by SDS-PAGE and analyzed forradioactivity. The ArcA protein (31 pmol) was phosphory-lated for 10 min with 14 pmol of 'ArcB in a 30-pl reactionmixture. Samples of 10 pl were treated as described above.We explored the kinetics of decay of 'ArcB-P and ArcA-P

at pH 12.5 by the following protocol. 'ArcB (30 pmol) wasphosphorylated for 2.5 min in 30 pl of the reaction mixture,and the protein was denatured by addition of 30 pl oftwice-concentrated loading buffer. Similarly, ArcA (31 pmol)was phosphorylated for 10 min by 14 pmol of 'ArcB in a30-pl reaction mixture, and the proteins were denatured.The two preparations were combined (120 p,l), prewarmedfor 2 min at 43°C, and alkalinized to pH 12.5 by addition of24 ,ul of 3 N NaOH. During this incubation, a 16-pl sample(originally containing 3.7 pmol of 'ArcB-32P and 2.6 pmolArcA-32P) was withdrawn at each interval and introducedinto a microcentrifuge tube containing 6.8 pl of 3 M MOPS toshift the pH back to 7.2. The samples were immediatelyfrozen in an ethanol bath cooled with dry ice and kept at thelow temperature until the last sample was taken. The pro-teins were separated by SDS-PAGE and analyzed for radio-activity.

RESULTS

Construction of plasmids pTB5 and pBA2 for the produc-tion of 'ArcB and ArcA. ArcB was found to be membraneassociated (20), and the deduced amino acid sequencesindicated that there were only two transmembrane se-quences near the N-terminal side. We therefore tried tosimplify our task of protein purification by constructing aplasmid bearing an arcB gene truncated from the 5' end sothat the 'ArcB product would lose the transmembrane se-quences and yet retain both the transmitter and receiverdomains. Previous studies on several sensors showed thatwhen the proteins were shortened, the remaining portionscontaining the transmitter domain were still able to undergoautophosphorylation and to act as a kinase of the cognateregulator protein (1, 5, 13, 21, 24). To construct a plasmid forthe production of the 'ArcB protein, pBB22 (19) containingarcB was digested by PstI and HindIII, and the 3.9-kbfragments were ligated to the vector pT7-7 (37) to give pTB5

'ArcB ArcA

A 1 2 3 B 1 2 3 4 5116.97-68'4329

,116-97'684329

FIG. 3. Purification of 'ArcB or ArcA. Samples from differentsteps of purification described in Materials and Methods weresubjected to SDS-PAGE (10% polyacrylamide gel) and then stainedwith Coomassie blue. (A) Results of 'ArcB purification. Lanes: 1,broken crude preparation of strain K38/pGl-2/pTB5 prepared bysonication (11 p.g of protein); 2, membrane fraction recovered afterultracentrifugation (11 ,ug of protein); 3, protein fraction recoveredfrom SDS-PAGE (5% acrylamide gel) (8 1lg of protein). (B) Resultsof ArcA purification. Lanes: 1, supernatant fraction after ultracen-trifugation of the crude extract prepared from strain K38/pGl-2/pBA2 (11 pg of protein); 2, fraction precipitated by ammoniumsulfate precipitation and dialyzed against the MOPS-EDTA buffer(10 p,g of protein); 3, eluate from DE52 column chromatography (12,ug of protein); 4, protein fraction recovered from SDS-PAGE (10%acrylamide gel) (10 p,g of protein); 5, combined fractions eluted by100 to 130 mM NaCl from the hydroxylapatite column (8 ,g ofprotein). Molecular weight markers are shown in unnumbered lanes;sizes are indicated in kilodaltons.

(Fig. 2). By this procedure, the PstI site (codon 129) in thearcB coding region deleted of the 5' segment encoding thetwo transmembrane portions was joined in frame down-stream of the vector translation site (codon 14) controlled bythe T7 promoter (Fig. 2).To construct a plasmid for the production of ArcA, pMW2

encoding arcA and phoM (17) was digested with KpnI andHindIII, and the 2.7-kb fragment containing arcA but notphoM was ligated to the vector pBluescript KS+ to givepBA2. By this procedure, arcA retaining its own promoterwas placed downstream of the T7 promoter (Fig. 2).

Purified 'ArcB. Because 'ArcB no longer possessed thetwo putative transmembrane segments on the amino-termi-nal side, we expected the protein to be mostly in thecytoplasmic fraction. However, when the truncated proteinwas overproduced by the T7 polymerase system, about halfof the total protein was associated with the membrane pellet.It is possible that the truncated protein tended to aggregateand was thus sedimented with the membrane fragments. Inthe case of truncated EnvZ, most of the protein was found inthe membrane fraction, and it was shown that this was theresult of aggregation and precipitation of the protein (13, 14).When purified 'ArcB was analyzed by SDS-PAGE (10%acrylamide gel), a single protein with an estimated molecularmass of about 70 kDa (close to the theoretical value of 74.5kDa) was found (Fig. 3). Such a protein was not present inpreparations from strain K38/pG1-2 with pBB22 bearing theintact arcB gene with a T7 promoter upstream or the hoststrain bearing the vector pT7-7 alone (data not shown).

Purified ArcA. When purified ArcA was fractionated in a10% acrylamide gel, a single protein with an estimatedmolecular mass of 25 kDa (close to the theoretical value of27.3 kDa) was found (Fig. 3). A smaller amount of theprotein was recovered from XL1-Blue cells bearing a plas-mid with the arcA coding region (cloned by polymerasechain reaction from the chromosome) placed downstream of

VOL. 174, 1992

on July 8, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

5620 IUCHI AND LIN

'ArcB 32pA1 2 3 B 1 2 3

FIG. 4. Autophosphorylation of 'ArcB. Each sample of 'ArcB(20 pmol) was incubated for 10 min with a different labeled ribonu-cleoside triphosphate: 0.1 mM [y-32]ATP (500 cpm/pmol) (lanes 1),0.1 mM [a- 2P]ATP (500 cpm/pmol) (lanes 2), or 0.1 mM [y-32P]GTP(500 cpm/pmol) (lanes 3). (A) Gel stained for protein; (B) 32pradioactivity associated with the proteins after exposure to X-rayfilm for 46 h.

the tac promoter of the vector pKK223-3 (Pharmacia). Nosuch protein was recovered from K38/pG1-2 cells bearingthe vector pBluescript KS+ (data not shown).

Autophosphorylation of 'ArcB. When purified 'ArcB wasincubated with [-y-32P]ATP, the electrophoretically sepa-rated, denatured, renatured protein became heavily labeled(Fig. 4). No labeling occurred when [a-32P]ATP was used,demonstrating that the protein was not simply binding to thesubstrate. Similarly, no significant labeling occurred when[_y-32P]GTP was used, demonstrating that the protein was

nucleoside triphosphate specific.As reported for the several sensor proteins (1, 9, 10, 39),

the phosphorylation of 'ArcB was rapid. The reactionreached the maximum within 2 min even at 4°C (data notshown), making it impractical to analyze the initial kinetics.The extent of 'ArcB autophosphorylation during 10 min ofincubation in reaction mixtures containing different concen-trations of [y-32P]ATP, however, indicated that 100 nM wasa feasible concentration to use for our further studies oftransphosphorylation from 'ArcB-P to ArcA (Fig. 5).

Phosphorylation of ArcA. When purified ArcA was incu-bated with [_y-32P]ATP for 5 min, the protein was not labeled(Fig. 6). However, the protein did become labeled when itwas introduced into a reaction mixture containing 'ArcBpreincubated for 2 min with [-y-32P]ATP. 'ArcB in thisprocedure was labeled to a lesser extent than that in thereaction mixture without addition of ArcA, which indicatesthat the radioactivity of ArcA was acquired at the expense ofArcB-32P. On the other hand, the radioactivity lost by'ArcB-32P was apparently less than the radioactivity ac-

F.

09

00

a

4c

_

1000-

800

600

400

200 I

100 200 300 400 500

32p labeled ATP (nM)

FIG. 5. Autophosphorylation of 'ArcB at various concentrationsof [-y-32P]ATP. Each sample of 'ArcB (20 pmol) was incubated for 10min with a different concentration of [-y- P]ATP (2,000 cpm/pmol).The radioactivity acquired by 'ArcB was determined with a Phos-phorimager (units given on the ordinate) and rendered visible byexposure to X-ray film for 17 h (inset).

A 1 2 3 Bi 2 3

'ArcB

ArcA

FIG. 6. Phosphorylation of ArcA by 'ArcB-32P. The Arc pro-teins were incubated in the reaction mixtures for various times.Lanes: 1, 'ArcB alone (9.8 pmol) was incubated for 7 min; 2, 'ArcB(9.8 pmol) was incubated for 2 min, ArcA (21 pmol) was then added,and the incubation was prolonged for 5 min; 3, ArcA alone (21 pmol)was incubated for 5 min. (A) Gel stained for protein; (B) 32pradioactivity associated with the proteins after exposure to X-rayfilm for 39 h.

quired by ArcA, indicating that further autophosphorylationof 'ArcB was occurring during the incubation.To monitor the time course of ArcA phosphorylation,

[_y-32P]ATP was added to a reaction mixture containing ArcAand 'ArcB at a molar ratio of 20:1. ArcA was phosphorylatedlinearly for at least 10 min (Fig. 7). The effect of 'ArcBconcentration on the rate of ArcA phosphorylation wastested during a 10-min incubation period. ArcA was added at21 pmol of ArcA, and 'ArcB was added from 0.15 to 9.8 pmol(Fig. 8). The linear relationship between the amount of'ArcB added to radioactivity acquired by purified ArcAindicated that only 'ArcB was limiting for the reaction.

Stability of the phosphate links of 'ArcB and ArcA. To testwhether both a His (presumably at position 292 in theconserved transmitter domain) and an Asp (presumably atposition 533 or 576 in the conserved receiver domain) of'ArcB are autophosphorylated and whether only an Asp(presumably at position 11 or 54 in the conserved receiverdomain) of ArcA is phosphorylated, we took advantage ofthe known acidic lability and the alkaline stability of thephosphate link to His in contrast to the instability of the

CD

I-

cc0

a

0 2 4 6 8 10 12TIME (min)

FIG. 7. Time course of phosphorylation of ArcA. A reactionmixture (50 ,u) containing ArcA (103 pmol) and 'ArcB (4.9 pmol)was incubated in the presence of [-y-32P]ATP. A 10-,uJ sample waswithdrawn at various time intervals. The radioactivity acquired byArcA was determined with a Phosphorimager (units given on theordinate) and rendered visible by exposure to X-ray film for 44 h(inset).

J. BACTERIOL.

0

on July 8, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHORYLATION OF Arc REGULATORY PROTEINS 5621

A 100

c 8

6I-

4 40al 2

B'ArcB

A

a

F.1-

0

0

1 0

'ArcB (pmol)

ArcA 4

FIG. 8. Phosphorylation rate of ArcA as a function of limitingconcentrations of 'ArcB. Each reaction mixture containng the sameconcentration of ArcA (21 pmol) but different concentrations of'ArcB (0.15, 0.3, 0.61, 1.22, 2.44, 4.9, or 9.8 pmol) was incubated for10 min. The radioactivity acquired by ArcA was determined with aPhosphorimager (units given on the ordinate) (A) and renderedvisible by exposure to X-ray film for 44 h (B).

30 40

TIME (min)

BArcB

ArcA

FIG. 10. Rate of alkaline hydrolysis of 'ArcB-32P and ArcA-32P.The mixture as described in Materials and Methods was incubated atpH 12.5 for various periods in a 43°C water bath. (A) Radioactivitiesretained by the proteins, determined with a Phosphorimager (unitsgiven on the ordinate); (B) radioactivities after 17 h of exposure in aPhosphorimager.

phosphate link to Asp under both conditions (5, 26). Whentested, 'ArcB-32P lost practically all of the label after incu-bation at pH 0.9 or pH 12.5 for 60 min but retained moderateamounts of the label at pH 2.7 or 7.2 (Fig. 9A). A similarpattern of label retention was shown by ArcA-32P (Fig. 9C).However, when examined by a Phosphorimager, the 'ArcBsample at pH 12.5 was found to retain 6% of the label.To determine whether the low amount of label retained by

'ArcB at pH 12.5 was attributable to a relatively stablelinkage, preparations of 'ArcB-32P and ArcA-32P denaturedat neutral pH were mixed and exposed to high pH. Samplesexposed for various times were analyzed for loss of label byeach protein. The 'ArcB protein showed a biphasic rate oflabel loss, whereas the ArcA protein showed only a rapidrate of label loss (Fig. 10). It therefore appears that 'ArcB-Pcomprised two kinds of phosphate link, whereas ArcA-Pcomprised only a single kind. A separate experiment carriedout to determine the ratio of the stable phosphate to labilephosphate before and after incubation of 'ArcB-P for 60 min

1 2 3 4 5

'ArcB

'ArcB B

ArcA C

FIG. 9. Stabilities of the phosphate links to 'ArcB-P and ArcA-P.(A) Radioactivities retained by autophosphorylated 'ArcB-32P afterincubation at different pH levels for 60 min at 43'C. Lanes: 1, pH 0.9treated; 2, pH 2.7 treated; 3, pH 7.2 treated; 4, pH 12.5 treated; 5,untreated control. (B and C) Results of a similar experiment inwhich the reaction mixture contained 'ArcB (B) and ArcA (C). TheX-ray film was exposed for 3 days.

at pH 12.5 indicated that about 8% of the phosphate linkageto ArcB was attributable to the stable phosphate link.

Physiological function of 'ArcB. Since arcB null mutantswere sensitive to the redox dye toluidine blue o (15, 19), wetested whether this mutation can be complemented by the'ArcB-producing plasmid pTB5. For this purpose, we con-structed an arcBl mutant of strain K38/pGl-2. One mutantwas transformed with pTB5, and the other was transformedwith the vector pT7-7. The transformants were spread ontryptone agar containing toluidine blue o and incubatedovernight at 37°C. Plasmid pTB5, but not the vector pT7-7,allowed growth of the arcBl mutant, although the colonysize of the strain with pTB5 was smaller than that of theparental arcB+ strain bearing the vector pT7-7. The reasonfor the difference in colony size is not clear, although it ispossible that growth retardation was caused by the overpro-duction of 'ArcB (Fig. 3). In any case, it is clear that 'ArcBconferred at least partial dye resistance.

DISCUSSIONSo far, five sensor proteins are known to possess both a

transmitter and a receiver domain: ArcB, BvgS, FrzE,RcsC, and VirA (3, 20, 35, 36). Autophosphorylation activitywas reported for VirA and FrzE. When the phosphorylatedVirA was analyzed, only histidyl phosphate was found in thealkaline hydrolysate. Under such an experimental condition,however, aspartyl phosphate would be too labile to berecovered (21). On the other hand, when FrzE was analyzed,it was noted only that the phosphate link under both acidicand alkaline conditions was unstable. Phosphorylation of theconserved His residue was inferred from homology of theregion to other sensor proteins (26). Our study on stability of'ArcB-P showed that the protein lost the phosphoryl groupat two different rates, suggestive of histidyl and aspartylphosphate links. We had previously found that replacementof the conserved His-292 by Gln or the conserved Asp-576 orAsp-533 by Ala prevented anaerobic repression of a targetoperon (19). Thus, phosphorylation probably occurred at the

Aimpmr

VOL. 174, 1992

on July 8, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

5622 IUCHI AND LIN

+ 02

ArcB

-02 -02

ArcA

-02 -02

ArcA

FIG. 11. Model for the signalling process by ArcB. Abbreviations: H, conserved His-292; D, conserved Asp-576 in ArcB or Asp-54 inArcA; P, phosphoryl group. Arrow drawn with broken line indicates stimulus; x indicates no signal transduction.

conserved His and one (or both) of the conserved Aspresidues. According to what is known about CheY phos-phorylation (31), the phosphorylated residue ofArcB is morelikely to be Asp-576.We also demonstrated the phosphorylation of ArcA by

'ArcB-P. The instability of the phosphate link to ArcA underextreme conditions of pH was indicative of an aspartylphosphate bond (probably at either the conserved Asp-11 orAsp-54). In this case, by deducing from CheY phosphoryla-tion (31), Asp 54-is more likely to be phosphorylated. Theproportionality between the rate of phosphorylation of puri-fied ArcA (in excess) and the concentration of 'ArcB purifiedis consistent with a direct intermolecular phosphoryl grouptransfer.The catalytic properties of 'ArcB support a proposed

model for the ArcB signal transduction process with thephenotypic consequences of replacing the conserved Hisand Asp residues ofArcB (19). Upon stimulation, the proteinfirst undergoes autophosphorylation at His-292. This phos-phoryl group is then transferred intramolecularly to a con-served Asp in the receiver domain. Following such a trans-fer, the liberated histidyl residue undergoes another round ofautophosphorylation. We conjecture that the phosphoryla-tion of ArcB at the conserved Asp changes the proteinconformation so that His-P of ArcB can now phosphorylateArcA (Fig. 11). On the other hand, we previously observedthat if ArcB was deprived of its receiver domain, theshortened protein could still transmit the signal, but theamplitude of target regulation between aerobic and anaero-bic conditions was reduced from 26- to 2-fold (19). Takentogether, the data indicate that autophosphorylation anddephosphorylation of the Asp in ArcB has an amplifyingeffect on the phosphorylation range of ArcA. Details of thismodel will be tested by analyzing the phosphorylating prop-erties of available mutant proteins.

ACKNOWLEDGMENTS

We thank S. Tabor for providing vector pT7-7, V. Rivera forinstruction in Phosphorimager usage, and A. Ulrich for help withpreparation of the manuscript.

This study was supported by Public Health Service grant R01-GM40993 from the National Institute of General Science.

REFERENCES1. Aiba, H., T. Mizuno, and S. Mizushima. 1989. Transfer of

phosphoryl group between two regulatory proteins involved inosmoregulatory expression of the ompF and ompC genes inEscherichia coli. J. Biol. Chem. 264:8563-8567.

2. Albright, L. M., E. Huala, and F. M. Ausubel. 1989. Prokaryoticsignal transduction mediated by sensor and regulator proteinpairs. Annu. Rev. Genet. 23:311-336.

3. Arico, B., J. F. Miller, C. Roy, S. Stibitz, D. Monackt S. Falkow,R. Gross, and R. Rappuoli. 1989. Sequences required for expres-sion of Bordetella pertussis virulence factors share homologywith prokaryotic signal transduction proteins. Proc. Natl. Acad.Sci. USA 86:6671-6675.

4. Drury, L. S., and R. S. Buxton. 1985. DNA sequence analysis ofthe dye gene of Escherichia coli reveals amino acid homologybetween the Dye and OmpR proteins. J. Biol. Chem. 260:4236-4242.

5. Forst, S., J. Delgado, and M. Inouye. 1989. Phosphorylation ofOmpR by the osmosensor EnvZ modulates expression of theompF and ompC genes in Escherichia coli. Proc. Natl. Acad.Sci. USA 86:6052-6056.

6. Fu, H.-A., S. Iuchi, and E. C. C. LUn. 1991. The requirement ofArcA and Fnr for peak expression of the cyd operon inEscherichia coli under microaerobic conditions. Mol. Gen.Genet. 226:209-213.

7. Gross, R., B. Arico, and R. Rappuoli. 1989. Families of bacterialsignal-transducing proteins. Mol. Microbiol. 3:1661-1667.

8. Hager, D. A., and R. R. Burgess. 1980. Elution of proteins fromsodium dodecyl sulfate-polyacrylamide gels, removal of sodiumdodecyl sulfate, and renaturation of enzymatic activity: resultswith sigma subunit of Escherichia coli RNA polymerase, wheatgerm DNA topoisomerase, and other enzymes. Anal. Biochem.109:76-86.

9. Hess, J. F., R. B. Bourret, and M. I. Simon. 1988. Histidinephosphorylation and phosphoryl group transfer in bacterialchemotaxis. Nature (London) 336:139-143.

10. Hess, J. F., K. Oosawa, N. Kaplan, and M. I. Simon. 1988.Phosphorylation of three proteins in the signalling pathway ofbacterial chemotaxis. Cell 53:79-87.

11. Hess, J. F., K. Oosawa, P. Matsumura, and M. I. Simon. 1987.Protein phosphorylation is involved in bacterial chemotaxis.Proc. Natl. Acad. Sci. USA 84:7609-7613.

J. IBACTERIOL.

on July 8, 2017 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHORYLATION OF Arc REGULATORY PROTEINS 5623

12. Huang, Y., P. Morel, B. Powell, and C. I. Kado. 1990. VirA, acoregulator of Ti-specified virulence genes, is phosphorylated invitro. J. Bacteriol. 172:1142-1144.

13. Igo, M. M., A. J. Ninfa, and T. J. Silhavy. 1989. A bacterialenvironmental sensor that functions as a protein kinase andstimulates transcriptional activation. Genes Dev. 3:598-605.

14. Igo, M. M., and T. J. Silhavy. 1988. EnvZ, a transmembraneenvironmental sensor of Escherichia coli K-12, is phosphory-lated in vitro. J. Bacteriol. 170:5971-5973.

15. Iuchi, S., D. C. Cameron, and E. C. C. Lin. 1989. A secondglobal regulator gene (arcB) mediating repression of enzymes inaerobic pathways of Escherichia coli. 1. Bacteriol. 171:868-873.

16. Iuchi, S., V. Chepuri, H.-A. Fu, R. B. Gennis, and E. C. C. Lin.

1990. Requirement for terminal cytochromes in generation ofthe aerobic signal for the arc regulatory system in Escherichiacoli: study utilizing deletions and lac fusions of cyo and cyd. J.Bacteriol. 172:6020-6025.

17. Iuchi, S., and E. C. C. Lin. 1988. arcA (dye), a global regulatorygene in Escherichia coli mediating repression of enzymes inaerobic pathways. Proc. Natl. Acad. Sci. USA 85:1888-1892.

18. luchi, S., and E. C. C. Lin. 1991. Adaptation of Escherichia colito respiratory conditions: regulation of gene expression. Cell66:5-7.

19. Iuchi, S., and E. C. C. Uin. 1992. Mutational analysis of signaltransduction by ArcB, a membrane sensor protein for anaerobicrepression of operons involved in the central aerobic pathwaysin Escherichia coli. J. Bacteriol. 174:3972-3980.

20. luchi, S., Z. Matsuda, T. Fujiwara, and E. C. C. LAn. 1990. ThearcB gene ofEscherichia coli encodes a sensor-regulator proteinfor anaerobic repression of the arc modulon. Mol. Microbiol.4:715-727.

21. Jin, S., T. Roitsch, R. G. Ankenbauer, M. P. Gordon, and E. W.Nester. 1990. The VirA protein ofAgrobacterium tumefaciens isautophosphorylated and is essential for vir gene regulation. J.Bacteriol. 172:525-530.

22. Keener, J., and S. Kustu. 1988. Protein kinase and phosphopro-tein phosphatase activities of nitrogen regulatory proteinsNTRB and NTRC of enteric bacteria: roles of the conservedamino-terminal domain of NTRC. Proc. Natl. Acad. Sci. USA85:4976-4980.

23. Kofoid, E. C., and J. S. Parkinson. 1988. Transmitter andreceiver modules in bacterial signaling proteins. Proc. Natl.Acad. Sci. USA 85:4981-4985.

24. Makino, K., H. Shinagawa, M. Amemura, T. Kawamoto, M.Yamada, and A. Nakata. 1989. Signal transduction in the phos-phate regulon of Escherichia coli involves phosphotransferbetween PhoR and PhoB proteins. J. Mol. Biol. 210:551-559.

25. Maniatis, T., E. F. Fritsch, and J. Sambrook 1982. Molecularcloning, a laboratory manual, Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

26. McCleary, W. R., and D. R. Zusman. 1990. Purification andcharacterization of the Myxococcus xanthus FrzE proteinshows that it has autophosphorylation activity. J. Bacteriol.172:6661-6668.

27. Ninfa, A. J., and R. L. Bennett. 1991. Identification of the site ofautophosphorylation of the bacterial protein kinase/phosphataseNRII. J. Biol. Chem. 266:6888-6893.

28. Ninfa, A. J., and B. Magasanik. 1986. Covalent modification oftheglnG product, NRI, by the glnL product, NRII, regulates thetranscription of the glnALG operon in Escherichia coli. Proc.Natl. Acad. Sci. USA 83:5909-5913.

29. Nixon, B. T., C. W. Ronson, and F. M. Ausubel. 1986. Two-component regulatory systems responsive to environmentalstimuli share strongly conserved domains with the nitrogenassimilation regulatory genes ntrB and ntrC. Proc. Natl. Acad.Sci. USA 83:7850-7854.

30. Ronson, C. W., B. T. Nixon, and F. M. Ausubel. 1987. Con-served domains in bacterial regulatory proteins that respond toenvironmental stimuli. Cell 49:579-581.

31. Sanders, D. A., B. L. Gillece-Castro, A. M. Stock, A. L.Burlingame, and D. E. Koshland. 1989. Identification of the siteof phosphorylation of the chemotaxis response regulator pro-tein, CheY. J. Biol. Chem. 264:21770-21778.

32. Silverman, P. 1986. Host cell-plasmid interactions in the expres-sion of DNA donor activity by F' strains of Escherichia coliK-12. Bioessays 2:254-259.

33. Silverman, P. M., S. Rother, and H. Gaudin. 1991. Arc and Sfrfunctions of the Escherichia coli K-12 arcA gene product aregenetically and physiologically separable. J. Bacteriol. 173:5648-5652.

34. Stibitz, S., and M.-S. Yang. 1991. Subcellular localization andimmunological detection of proteins encoded by the vir locus ofBordetella pertussis. J. Bacteriol. 173:4288-4296.

35. Stock, J., A. J. Ninfa, and A. M. Stock. 1989. Protein phosphor-ylation and regulation of adaptive responses in bacteria. Micro-biol. Rev. 53:450-490.

36. Stout, V., and S. Gottesman. 1990. RcsB and RcsC: a two-component regulator of capsule synthesis in Escherichia coli. J.Bacteriol. 172:659-669.

37. Tabor, S., and C. C. Richardson. 1985. A bacteriophage T7RNA polymerase/promoter system for controlled exclusiveexpression of specific genes. Proc. Natl. Acad. Sci. USA82:1074-1078.

38. Weber, R. F., and P. M. Silverman. 1988. The Cpx proteins ofEscherichia coli K12. Structure of the CpxA polypeptide as aninner membrane component. J. Mol. Biol. 203:467-478.

39. Weiss, V., and B. Magasanik. 1988. Phosphorylation of nitrogenregulator I (NRI) of Escherichia coli. Proc. Natl. Acad. Sci.USA 85:8919-8923.

VOL. 174, 1992

on July 8, 2017 by guesthttp://jb.asm

.org/D

ownloaded from