Signal Transduction and Osmoregulation in Escherichia coli

Vol. 266, No. 17, Issue of June 15, pp. 10775-10780,1991 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY

(0 1991 by The American Society for Biochemistry and Molecular Biology, Inc

Signal Transduction and Osmoregulationin Escherichia coli ANOVEL TYPE OF MUTATION IN THE PHOSPHORYLATION DOMAIN OF T H E ACTIVATORPROTEIN, OmpR,RESULTSIN A DEFECTINITSPHOSPHORYLATION-DEPENDENTDNABINDING* (Received for publication, November 27, 1990)

Kyoko Nakashima, Kengo Kanamaru, HirofumiAiba, and Takeshi MizunoS From the Laboratory of Microbiology, School of Agriculture, Nagoya Uniuersity, Chikusa-ku, Nagoya 464, Japan

The transcriptional factors, OmpR and EnvZ, are crucially involved in the osmotic regulation of o m p F and o m p C expression in Escherichia coli. The DNA binding ability of the positive regulator, OmpR, is modulated through its phosphorylation and dephosphorylation mediated byEnvZ in response to the medium osmolarity. In this study, two examples of a novel type of mutant o m p R allele, o m p R 9 6 A and ompR115S, whose phenotype is OmpF- OmpC- irrespective of the medium osmolarity, were characterized. These mutations result in amino acid conversions, G ~ touAla~ and ~ Arg’’’ to Ser, respectively, withinthe phosphorylation domain of OmpR. Nevertheless, these mutant proteins were capable of undergoing phosphorylation and dephosphorylation normally, just like wild-type OmpR. However, the phosphorylation-dependent enhancement of their in vitro DNA binding ability was found to be severely affected. It was thus revealed that these mutant OmpR represent a novel type in terms of the mechanism ofphosphorylation-dependentactivation of the function of OmpR, i.e. those are normally phosphorylated but not activated to bind to the cognate promoter DNAs. In this respect, it was further suggested that OmpR oligomerization may be involved in the mechanism underlying the phosphorylation-dependent enhancement of the DNA binding ability of OmpR. The mutant proteins characterized in thisstudy seem to be defective in this particular oligomerization process observed in vitro.

phorylation of OmpR results in enhancement of not only its DNA binding ability as to both ompF the and ompC promoter DNAs but also the transcription of these genes ( 5 , 6). The mechanistic question then arose as to how the phosphorylation of OmpR influences its DNA binding ability. In this study, to address this mechanistic question,two examples of a novel type of mutant OmpR proteinwere characterized. Many different types of ompR mutants have been isolated and characterized at themolecular level in terms of not only their i n vivo osmoregulatory phenotypes but also the i n vitro biochemical properties of their gene products (1,7-11). Based on these results, itwas suggested that the N-terminalhalf of the OmpR molecule contains a site involved in phosphorylation by EnvZ, whereas the C-terminalhalf possesses a DNAbinding sitefor the ompF and ompC promoter DNAs (12,13). During the course of our previous genetic studies, we happened to isolatetwo independent ompR mutants,both of which exhibit an OmpF- OmpC- phenotype irrespective of the medium osmolarity. At that time, however, the biochemical properties of thesemutantOmpRproteins were not assessed further. In the light of recent findings as to EnvZOmpR phosphotransfer (14-16), we recalled these proteins and characterized themin this study in termsof their EnvZOmpR phosphotransfer reactions and DNA binding properties. The results obtained in this study on the biochemical properties of these mutant OmpR proteins, each of which has a single amino acid conversion within the N-terminal phosphorylation domain,provided us witha clue for understanding the mechanism by which OmpR phosphorylation influences the DNA binding abilityof OmpR.

Expression of the Escherichia coli outer membrane proteins, EXPERIMENTALPROCEDURES OmpF and OmpC, is regulated in response to the medium osmolarity. The transcriptional factors, OmpR and EnvZ, are Materials-Restrictionendonucleases, the Klenow fragment of kit, and aDNAsequencing kit crucially involved in theosmotic regulationof ompF and ompC DNA polymerase,aDNAligation expression, namely OmpRis the actual activator,which binds were purchased from Takara Shuzo Co. Ltd. [y-”P]ATP (30 Ci/mmol) and [a-:”P]CTP (3000 Ci/mmol) were to both theompF and ompC promoter DNAs, and EnvZ is a obtained from Amersham Corp. The protein cross-linking reagent, transmembrane osmotic sensor, which exhibits both OmpR dimethyl suberimidatedihydrochloride (DMS),’ was from Wako Pure phosphorylation and dephosphorylation abilities (see Refs. 1- Chemical. All other materials were of reagent grade. 4 for reviews). Recently, itwas revealed that phosphotransfer BacterialStrainsand Plasmids-E. coli K12 strain SG480A76 between thesetwo regulatory components, originally observed (AmalT-ompB AlacU169 araD rpsL relA thiA flbB) was used as a host i n vitro, appears to play a crucial role in signal transduction strain for determination of the osmoregulatory phenotypes(17). and the consequent osmotic regulation i n vivo (see Refs. 3 Strains SG480A76 and AT142 (AenuZ AlacU169 araD rpsL re& thiA flbB) were used as host strains for purification of the OmpR and and 4 for review). In this respect,we previously addressed the EnvZproteins, respectively (18). Asingle copy numberplasmid, main questionof what thebiochemical consequence of OmpR pMAN104, which carries the wild-type ompR and enuZ genes, was phosphorylation is and demonstratedi n vitro that the phos- constructed previously (19). Plasmids carrying the mutantompR96A * This work was supported by grants from the Ministryof Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ T o whom correspondenceshould be addressed.Tel.: 052-7815111 (ext. 6289); Fax: 052-781-0693.

and ompR115S genes connected with the wild-type enuZ gene were constructed from pMAN104 by replacing the ompR region with the novel ompR mutant alleles (plasmids pNAKOOl and pNAKOO2, respectively). 1 The abbreviations used are: DMS, dimethyl suberimidate dihydrochloride; SDS, sodium dodecyl sulfate; AMP-PNP, adenosine 5’P,-imin0)triphosphate.

10775

10'776

in E. coli

Activation Osmoregulation Gene and

DNA Techniques-The conditions used for DNA-manipulating enzymes, such as restriction endonucleases and a DNA ligation kit, were those recommended by the supplier. DNA sequences were determined by the dideoxy chain termination method, usinga commercially available sequencing kit (20). Purification of the OmpR and EnuZ Proteins-The wild-type and mutant OmpR proteins used in this studywere purified as described previously (21) with slight modifications and with fast protein liquids Mono-Qchromatography(PharmaciaLKB Biotechnology,Inc.) being performed as the final stepof purification. A truncated form of the EnvZ protein (termed hereafter as EnvZ*),which comprises the amino acid residues from Tyr"' to C-terminal Gly"", waspurified fromwild-type and mutant (enuZ11) cells as describedpreviously (15). Immunoblot Ana/.ysis-Protein samples weresubjected to SDSpolyacrylamide gel (12.5% acrylamide) electrophoresis. Proteins on the gels were transferred to nitrocellulose filters as described previously (22). The filters were treated with anti-OmpR and anti-EnvZ antisera and then treated with alkaline phosphatase-coupled goat anti-rabbit immunoglobulin G to detect cross-reacting proteins. PhosphorylationExperiments-The in uitro experimental conditions for EnvZ aut.ophosphorylation, phosphotransfer between EnvZ* and OmpR, and OmpR dephosphorylationwere essentially the same as those given previously (9, 15). Abuffer named TEDG (50 mM Tris-HCI, pH 8.0, 0.5 mM EDTA, 2 mM dithiothreitol, 10% glycerol (v/v)) was mainly used in these experiments. Purification of radioactively prephosphorylated EnvZ* was carried out as described previously (15). The sampleswere analyzed by SDS-polyacrylamide gel electrophoresis (12.5% acrylamide) (23) followed by autoradiography with Fuji x-ray film (RX-50). DNA Binding Assay-The DNA binding with OmpR was assayed bymeans of nondenaturing gel retardationanalysis,as described previously (21). Preparation of Outer Membrane Proteins-Outer membrane proteins were prepared by extraction of cell envelopes with sodium N lauroyl sarcosinate and then analyzed by urea-SDS-polyacrylamide gel electrophoresis as described previously (18). Cross-linking Experiments-Purified OmpR was incubated with EnvZ11' in TEDG buffer containing 400 mM KC1 and 5 mM MgCI, in the presence of 0.1 mM ATP for 30 min at 37 'C. Each sample was subjected to cross-linking with a protein cross-linker, DMS, as describedpreviously (24). Briefly, after adding triethanolamine and DMS to concentrations of 200 and 40 mM, respectively, each reaction mixture was kept for 1 h a t room temperature. Proteins were precipitated by adding 4-fold in a volume of cold ethanol and then analyzed by SDS-polyacrylamide gel (10% acrylamide) electrophoresis.

phosphotransfer, however, we recalled these mutant ompR genes and determined their nucleotide sequences inthis study. It was revealedthat both mutantompR genes have single base substitution that results in amino acid conversion within the N-terminal phosphorylation domain of OmpR, as shown in Fig. 1. One results in a Glu to Ala conversion at the 96th amino acid of OmpR and the other in an Arg to Serconversion at the 115th amino acid. These mutant ompR alleles were thus designated as ompR96A and ompRl15S, respectively. It seemed to us that these ompR mutants are examples of a novel type of ompR mutant as compared with a variety of ompR mutantscharacterized previously(see Fig. 1). This findingpromptedustocharacterizethesemutant OmpR proteins in terms of their biochemical properties in vitro. I n Vivo Osmoregulatory Phenotypes of the ompR MutantsFirst of all, to clarify the in vivo osmoregulatory phenotypes exhibited by the mutant ompR alleles, ompR96A and ompR115S, they were connected with the wild-type envZ gene on a single copy number plasmid. The resultant plasmids were designatedas pNAKOOl and pNAKOO2, respectively. An ompR-envZ deletion strain, SG480A.76,was transformed with these plasmids as well as a plasmid, pMAN104, carrying the wild-type ompR and envZ genes. Outer membrane proteins were prepared from cells grown in a medium supplemented with and without 20% sucrose and then analyzed by ureaSDS-polyacrylamide gel electrophoresis (Fig. 2). Cells carrying the wild-type ompR gene displayed the normal osmoregulatory phenotype as expected (see p M A N I 0 4 ) , while amounts of both the OmpF and ompC proteinsproduced in ,

Phosphorylalton

DNA-Btndmg

~q~~~G NH 2'

- -

$ Cys (F- C')

*

I

*I

-

A&?

G(u96A!415

Gln (F-C-)

Ala S Q ~ (F-CjF-C7

~

~

u

AGl,%r$2k03

(F-C-)

~

-

AIZm

kOOH

Met

(F' C-)

FIG. 1. Schematicrepresentation of thestructure of the OmpR protein and summary of the amino acid changesin the mutant OmpR proteins. The open and shaded rectangles denote the phosphorylation and DNA-binding domains in the OmpR protein, respectively. The positionsof amino acid residues are numheredfrom the N terminus of the protein. Four different types of mutant OmpR RESULTS proteins,each of which haseitheranamino acid conversion or Isolation and Characterization of Examples of a Novel Type deletion, have been characterized previously (1, 7-11). Two examples of ompR Mutant-The mutant ompR genes characterized in of a novel type of mutant OmpR proteinwith an aminoacid converthis study were originally isolated from a strain carrying the sion (GI@ to Ala and Arg"" to ser, respectively) at the position mutant envZl1 allele on its chromosome, which also harbors indicated by asterisks were characterized in this study. The osmorephenotype of cells producing each mutant OmpR proteinis a single copy number plasmid carrying the wild-type ompR gulatory indicated in parentheses in which F and C' denote the production and mutant envZll genes. Thus, this particular strain is a of the OmpF and OmpC proteins in the outer membrane, respectively.

mero-diploid with respect to thewild-type ompR and mutant envZl1 genes. The envZl1 mutation is known to exhibit an OmpF- OmpC-constitutive phenotype (25). This particular missense mutation is also known to be pleiotropic, affecting the synthesisof a subset of proteins, including alkaline phosphatase and proteinsinvolved in maltose utilization, through a yet unknown mechanism (26). Thus, the enuZll mutant is phenotypically OmpF- Mal- PhoA-. Starting from this merodiploid strain, we previously attempted to isolate mutants capable of growing on a maltose minimal medium, and two independent mutants exhibiting the Mal+ phenotype were obtained. In both mutants, a mutation conferring the Mal+ phenotype turned out to be located within the ompR gene carried on the plasmid used (data not shown). However, we ompR gene fail to then found that these mutations in the confer the O m p F phenotype on the mero-diploid strain in which we were much interested. Therefore,at thattime, these mutant ompR genes were not subjected to further characterization. In the light of recentfindings as to EnvZ-OmpR

Strain SG480&76(AompB) @-4F21 pM4NIOLpNpKoo1pNAKoo2

n n n n

C""".-F -

0 2 0 0 2 0 0 2 0 0 2 0 Sucrose(%)

FIG. 2. Urea-SDS-polyacrylamide gel electrophoresis showing expression of the ompF and ompC genes. Strain SG480A76, which is an ompR (ompH-enu%)deletion mutant, was transformed with each single copy number plasmid: pMF21 (a vector plasmid), pMAN104 carrying the wild-type ompii and enuZ genes, pNAKOOl carrying the mutant ompR96A and wild-type enuZ genes, pNAK002 carrying the mutantompR115S and wild-type e n d genes. The cells were g o w n in a medium containing the indicated concentrations of sucrose. Outermembraneproteins were prepared and analyzed by urea-SDS-polyacrylamide gel electrophoresis. The positions of the OmpC and OmpF proteins areindicated by letters C and F, respectively.

Gene Activation and Osmoregulation in E. coli cells carrying the ompR96A and o m p R l l 5 S genes,respectively, were greatly reduced irrespective of the medium osmolarity (see pNAKOOl and pNAKOO2). T o exclude the possibility that these mutant OmpR proteins in cells are extremely unstable, total cell lysates were prepared from cells carrying these mutant genes and then subjected to immunoblot analysis with the use of an anti-OmpR antiserum (Fig. 3). The results of immunoblot analysis confirmed that the amounts of the respective mutant OmpR proteins in cells are comparable with that of the wild-type OmpR protein. We thus concluded that these ompR mutations, which result in a single amino acidconversion within the N-terminal phosphorylationdomain, specify a functionallyalteredOmpR protein and confer on cells the striking phenotype, OmpFOmpC-. PhosphorylationandDephosphorylation of theMutant OmpR Proteins-Based on the phenotype of the ompR96A and ompR115S mutants, we suspected that their gene products are mostlikely defective in an essential OmpR function. First of all, the phosphorylation and dephosphorylation of these mutant OmpR proteinswere examined i n vitro (Fig. 4). The wild-type and mutant OmpR proteins were purified by essentially the same strategies as described previously (21). The proteins thus purified from theompR96A and ompR115S mutant cells were designated as OmpR96A and OmpR115S, respectively. On the other hand, a truncated form of the EnvZ protein (EnvZ*), which was well characterized previously (9, 15), was also purified and incubated with0.1 mM [-y-””P]ATP. Radioactively prephosphorylated EnvZ* waspurifiedfree (lane C) andthenincubatedwithmutants fromATP OmpR96A and OmpR115S as well as wild-type OmpR in the absence of ATP in a buffer containing 5 mM MgC12 and 50 mMKC1 for 30 s a t room temperature. The samples were analyzed by SDS-polyacrylamide gel electrophoresis followed by autoradiography. An autoradiogram of the gel showed that mutants OmpR96A andOmpRll5Sas well as wild-type OmpR were efficiently phosphorylated, as shown in Fig. 4. (lanes 1 ). The addition of 0.1 mM cold ATP to the reaction mixtures resulted in acceleration of the dephosphorylationof

10777

OmpR96A and OmpR115S as well as wild-type OmpR (lanes 2-4). We also used a mutant EnvZ protein, namedEnvZ11*,

as an alternative phosphate donor for i n vitro OmpR phosphorylation (9). EnvZ11* is known to be capable of phosphorylat.ing wild-type OmpR but not tobe capable of mediating OmpR dephosphorylation. Essentially the samein vitro phenomena observed for EnvZ11* and wild-type OmpR were also observed for EnvZ11* and the mutant OmpR proteins (data not shown). We thusconcluded that as far asin vitro OmpR phosphorylation and dephosphorylation mediated by EnvZ are concerned, these mutant OmpR proteins are indistinguishable from wild-type OmpR, even though theformer each haveanamino acid substitution in theN-terminal phosphorylation domain. Phosphorylation-dependent Enhancement of theDNA Binding Abilityof the Mutant OmpR Protein-Thefunctional role of OmpR phosphorylation was assessed previously, and it was demonstrated that the phosphorylation of OmpR in its N-terminal domain enhances the ability of the C-terminal (13). We examined such domain to bind to the cognate DNAs phosphorylation-dependent enhancement of the DNA binding ability of mutants OmpR95A and OmpR115S (Fig. 5). In theseexperiments, a DNA fragment end-labeledwith IV2P, which encompasses either the ompF or ompC promoter sequence, was mixed with mutants OmpR96A and OmpR115S as well as wild-type OmpR. Various amounts of EnvZ11*, whichwasradioactively prephosphorylated with “‘P, were further added to the mixtures to phosphorylate OmpR fol, OmpR OmpR96AOmpRll5S 1 2 3 6 1 2 3 L 1 2 3 L ~

-0mpR-P

OmpR96A

1

2

3

4

r

1

2

3

L

---OOmpR

FIG. 3. Immunoblot detection of the OmpR protein. Strain SG480A76, which is an ompR ( o m p R - e n d ) deletion mutant, was transformed with each single copy number plasmid:lane 1 , pMF21 (a vector plasmid); lane 2, pMAN104 carrying the wild-type ompR and enuZ genes; lane 3, pNAKOOl carrying the mutant ompR96A and wild-type enuZ genes; lane 4 , pNAKOO2 carrying the mutant ompRZZ5S and wild-type enuZ genes. The cells were grown in Luria broth. Total cell lysates were prepared and subjected to immunoblot analysis with the use of an anti-0mpR antiserum.

d

l

/

RQlallveAmount Of Phosphoryiatrd OmpR

FIG. 5. Phosphorylation-dependent enhancement of the

DNA binding ability of the OmpR protein as to the ompC and ompF promoter DNAs and its quantitative representation. A c DNA fragment end-labeled with nrP, which encompasses either the EnvZ? ompC or ompF promoter sequence (1.7 and 1.5 fmol, respectively), was mixed with wild-type OmpR, OmpR96A, or OmpR115S (20 pmol -om*-P each).To each mixture, various amounts of radioactively phosphorylh2P n“rP Ai”P ated EnvZ11* (lane I, 0 pmol; lane 2, 2 pmol; lane 3 , 10 pmol; lane 4 , FIG. 4. Autoradiogram showing the time course of OmpR 50 pmol) were added and incubated in 40 p1 of TEDG buffer containphosphorylation and dephosphorylation mediated by the ing 5 mM MgCL and 400 mM KC1 for 10 min at room temperature. EnvZ protein. Radioactively phosphorylated EnvZ* (40 pmol) was Each sample was divided into two. One half was analyzed hy SDSprepared as described previously (lane C) and incubated with wild- polyacrylamide gel electrophoresis followed by autoradiography type OmpR, OmpR96A, or OmpR115S (60 pmol each) in 40 pl of (panel A ) . The other half was subjected to an in vitro DNA hinding TEDG buffer containing 5 mM MgCI, and50 mM KC1 for 30 s a t assay by means of nondenaturing gel electrophoresis followed by room temperature (lanes I). Subsequently,cold ATP (final 0.1 mM) autoradiography.The relative amounts of an OmpR-DNA complex was added at the indicated time (30 s ) and incubated further for 1 formed under these conditions were determined by scanning the min (lanes 2 ) , 2 min (lanes 3),and 4 min (lanes 4 ) . Aliquots were autoradiogram and plotted as a function of the relative amounts of removed from the reaction mixtures and subjectedto SDS-polyacryl- phosphorylated OmpR (panels R and C), which were calculated on amide gel electrophoresis followed by autoradiography. the basis of the autoradiogram shown in panel A . OmpR9&4O~~S 1 2 3 6 1 2 3 4 1 2 3 1

10778

Activation Osmoregulation Gene and

in E. coli

lowed by incubation for 10 min at room temperature. Each of exists as a monomeric form in solution (21). To address the these sampleswas divided into two. One half was analyzed by mechanisticquestionsas to how thephosphorylation of SDS-polyacrylamide gel electrophoresis (Fig. 5A), and the OmpR influences its DNA binding ability and why the phosother half was assayed as to in vitro DNA binding by means phorylated formsof mutants OmpR96A and OmpR115S charof nondenaturing gel retardation analysis (Fig. 5, B and C). acterized inthisstudyare defectivein the DNA binding As shown in Fig. 5A, the relative amount of the phosphoryl- ability, we attemptedto reassess the possibility that the ated form of OmpR in the reaction mixtures increased in phosphorylated form of OmpR may exist as an oligomeric proportion to the amount of the prephosphorylatedEnvZ11* form by means of an i n uitro cross-linking experiment. The added (lunes 1-4). The same sampleswere also assayed as to phosphorylated form of OmpR was prepared in uitro by inin vitro DNA binding by means of nondenaturing gel retar- cubating wild-type OmpR with EnvZ11* in the presence of dation analysis, followed by autoradiography. As shownin 0.1 mM ATP. The reaction mixturewas treated with a crossFig. 5, R and C, the relative amount of anOmpR-DNA linking reagent, DMS, at the final concentration of 40 mM complex formed undertheseexperimentalconditions was for 1 h a t room temperature and then analyzed by SDSdetermined by scanning the autoradiogram densitometrically polyacrylamide gel electrophoresis followed by staining with andplottedas a function of the relative amount of the Coomassie Brilliant Blue. As shown in Fig. 7, the stained gel phosphorylated form of OmpR in the reaction mixture. The showed that in addition to bands corresponding to EnvZ11* results of these experiments are shown in Fig. 5R (for the and OmpR, two additional stained bands with lower electroompC promoter) andC (for theompF promoter). It was clearly phoretic mobilities appeared (lune 3 ) .When OmpR alone was demonstrated thatin marked contrast to the case of wild-type incubated with DMS in the presence of ATP, such extra OmpR, phosphorylation-dependent enhancement of the i n bands were not observed (lane 4 ) asreported previously. uitro DNA binding ability as to both the ompF and ompC Furthermore, even when OmpR was incubated with EnvZ11* promoter DNAs was severely affected in the cases of mutant intheabsence of ATP (lune I ) or in the presence of a nonhydrolyzable ATP analog, AMP-PNP (lane Z ) , no addiOmpR96A and OmpR115S. It should be noted here that we previously reported that tionalstainedbands appeared. Theapparent molecular the nonphosphorylatedform of purified OmpR is also capable weights estimated for the extra bands on thegel were 55,000 of binding to theompC and ompF promoter DNAs, albeit with and 78,000, which are in good agreement with integral mula lower efficiency under certain conditions, i.e. when a high tiples of the apparent molecular weight determined for the concentration of OmpR was used for the DNA binding assay OmpR monomer (26,500). These results suggested that the (5, 21). As shown in Fig. 6, the nonphosphorylated forms of phosphorylation event on OmpR molecules may induce an mutants OmpR96A and OmpR115S as well as wild-type oligomeric interaction between them i n uitro. OmpR were capable of binding to the cognate DNAs to nearlyMutants OmpR96A and OmpR115S were subjected to the the same extent, provided that a high concentration of OmpR same cross-linking experiment in comparison with wild-type was used for the DNA binding assay. This suggested that a OmpR as shownin Fig. 8. These proteinswere incubated with latent DNA binding ability carried in the C-terminal domain EnvZ11* in the presence and absence of 0.1 mM ATP and of OmpR is not impaired, even in the cases of the mutant then treated with DMS. Each sample was divided into three proteins. Taking all these results together,we concluded that portions,each of whichwasapplied on a separateSDSmutants OmpR96A and OmpR115S are defective in a specific polyacrylamide gel. It should be noted here that under these OmpR function, i.e. phosphorylation-dependent efficient rec- particular experimental conditionsused, the mutant proteins ognition of the ompF and ompC promoter DNAs. were phosphorylated to nearly the same extent as that obOligomerization of the OmpR Protein Was Induced in Vitro servedfor the wild-type protein(datanotshown). After through Phosphorylation-We previously reported that puri- electrophoresis, one of the gels was stained with Coomassie fied OmpR, which is presumably the nonphosphorylated form, Brilliant Blue (panel A ) . The other two were subjected to immunoblot analysis with the use of either an anti-OmpR antiserum (panel B ) or an anti-EnvZ antiserum (panel C). 1 2 3

4 4

-

- 4

--

-EnvZ11

*

'm- m-OrnpR FIG. 6. Binding of the nonphosphorylated form of the OmpR protein to the ompC and ompF promoter DNAs. A DNA fragment (pmol) encompassing either the ompC or ompF promoter sequence (O..i8 and 0.65 pmol,respectively)was mixed withvarious amounts of wild-type OmpR, mutant OmpR96A, or mutant OmpR115S (lane I, 22 pmol; lane 2, 44 pmol; lane 3,88 pmol) in 10 pl of TEDG buffer containing 5 mM MgCI, and 400 mM KC1 and then incubatedfor10min at room temperature.Thesesampleswere subjected to in vitro DNA binding assay by means of nondenaturing gel electrophoresis followed by staining with ethidium bromide.Lane C represents the free DNA fragments encompassing the respective ompC and ompF promoter sequences.

FIG. 7. Phosphorylation-dependent oligomerization of OmpR protein observed in vitro. Wild-type OmpR (369 pmol) was incubated with EnvZ11* (123 pmol) for 30 min a t 37 "C in 25 pI of a buffer containing 6 mM MgCI, and 400 mM KC1 in the absence of ATP (lane 1 ) or in the presence of 0.1 mM AMP-PNP (lane 2 ) or 0.1 mM ATP (lane 3 ) . Eachmixturewastreatedwiththecrosslinking reagent, DMS, at thefinal concentration of 40 mM for 1 h in the presence of 200 mM triethanolamine and then analyzed by SDSpolyacrylamide gel electrophoresis followed by staining with COOmassie BrilliantHlue. The same cross-linking experiment was carried out for OmpR alone in the presence of 0.1 mM ATP without the addition of EnvZ11' (lane 4 ) . Triangles indicate the bands, which presumably correspond to OmpRoligomers.

Gene Activation and Osmoregulation in E. coli A

B

C

10779

enhancement of its DNA binding ability (5). The main mechanisticquestionasto how thephosphorylation of OmpR influences its DNA binding ability has not yet been answered. \ In this study, to address this mechanistic question, we isolated and characterized two samples of a novel type of mutant ompR allele, demonstratingthatthesemutantproteins OmpR96A and OmpR115S, are defective in a specific OmpR " "W function, i.e. phosphorylation-dependent efficient recognition of the ompF and ompC promoters. The results of further in "Y vitro characterization of these particular mutant proteinsby lmmunoblot lmmunoblot means of a cross-linkingexperiment provided us with an (antl-0mpR) (antl-EnvZ) mechanismunderlying the FIG. 8. Mutants OmpR96A and OmpRl15S are defective in implication tounderstandthe phosphorylation-dependent OmpR oligomerization. Wild-type phosphorylation-dependent enhancement of the DNA bindOmpR (lane I ) , OmpR96A (lane 2), or OmpRll5S ( h e 3 ) (369 pmol ing ability of OmpR. each) was incubated with EnvZ11' (123 pmol) in the absence (lanes A variety of missense and deletion mutations of the ompR denoted by -) and presence (lanes denoted by +) of 0.1 mM ATP. gene havebeenisolated that show altered osmoregulatory These samples were subjected to in uitro cross-linking under exactly the same conditions given in the legend to Fig. 7. Each sample was phenotypes under different osmotic conditions (1, 7-11). In divided into three portions and then applied on a separate SDS- this study, two examples of a novel type of mutant ompR polyacrylamide gel. After electrophoresis, one of the gels was stained allele whose phenotype is OmpF- OmpC- irrespective of the by Coomassie Brilliant Blue (lane A ) , and the other two were sub- medium osmolarity were characterized in terms of the in vitro jected to immunoblot analyses with the use of either an anti-OmpR biochemical properties of their gene products. The ompR96A antiserum (pane[ R ) or an anti-EnvZ antiserum (panel C). Arrowand ompR11SS alleles characterized in this study result in a heads indicate the bands, which presumably correspond to OmpR single amino acid conversion, Glu"; to Ala and Arg"" to Ser, oligomers. respectively, within the phosphorylation domain of OmpR. Based onthephenotype of the ompR96A and ompRl1SS Several intriguing resultswere obtained as follows. First, the gene products aremost likely extra bands, which were newly generated only upon incuba- mutants, we suspected that their tion of wild-type OmpRwith EnvZ11* in the presence of ATP defective in an essential OmpR function, e.g. OmpR phosphorylation or DNA binding ability, bothof which are essen(see lune I ) , were cross-reacted exclusively with the antitially involved in the mechanism underlying the activationof OmpRantiserum, i.e. notwiththeanti-EnvZantiserum. EnvZ11* used in this study appeared to be cross-linked to a ompF and ompC transcription. It was revealed that even small extent, regardless of the existence of ATP as well as though these mutant proteins were capable of undergoing in normally, phosOmpR in the reaction mixtures. This is probably due to the vitro phosphorylation and dephosphorylation fact that the truncated form of EnvZ in a solution tends to phorylation-dependent enhancement of their in vitro DNA aggregate (15). Furthermore, the amount of the cross-linked binding ability was severely affected, i.e. even when phosband(s) was varied in experiments and thus notreproducible phorylated they were not capable of binding efficiently to either the ompF or ompC promoter DNA. These in vitro (data not shown). In any event, these results further conobservations can simply account for the in vivo osmoregulafirmed that the newly generated bands correspond to phosphorylation-dependent oligomeric forms of OmpR (dimer and toryphenotype, OmpF- OmpC-, exhibited by these ompR trimer). It should be noted that a faint band probably corre- mutants. The paradoxical question then arose as to how the amino sponding to an OmpR tetramerwas also detected on immunoblotting with the use of the anti-OmpR antiserum (panel acid substitution, which occurred withinthe N-terminalphosB, lune 1). More importantly, in marked contrast to thecase phorylation domain of ompR, influences the DNA binding affecting the of wild-type OmpR, such presumable OmpR oligomerization ability carriedin the C-terminal domain without was not significantlyobservedfor mutant OmpR96A and OmpR phosphorylation event itself. In this respect, at least, OmpRllEiS (lunes 2 and 3 ) . These results may suggest that two mechanismscan be considered as follows. First,the the mutant proteins are defective in OmpR oligomerization, OmpR phosphorylation eventmay primarily induce a conforwhich was demonstrated in vitro to beinduced by OmpR mational change of the N-terminal portion of OmpR and subsequently may induce a conformational change of the Cphosphorylation. terminal portionin such a way that allosteric communication DISCUSSION between them occurs, which would lead to a change in its The results of previous genetic studies support the view DNA binding ability.Second, OmpR phosphorylation may that theOmpR protein is composed of at least two functional simply induce oligomerization of OmpR molecules and condomains, namely one is responsible for the interaction with sequently may induce apossible interaction between adjacent the EnvZ protein,whereas the other participatesin the inter- C-terminal portions of OmpR molecules, which would lead to action with both theompF and ompC promoter DNAs (7,27). a change in its DNA binding ability. The former possibility is The results of biochemical studies confirmed that these do- rather difficult to assess by means of only simple biochemical mains of OmpR are physically and functionally separable (see experiments. Here,we attempted to assess the latterpossibilFig. 1) and that the N-terminal half comprising the amino ityby means of an in vitro cross-linking experiment. We acid residues from N-terminal Met to Arg"" contains a site previously reported thatpurified OmpR, which is presumably involved in phosphorylationby EnvZ (12,13).It hasalso been the nonphosphorylated form, exists as a monomeric form in of cross-linkingexperconfirmed that the C-terminal half comprising the amino acid solution (21). In this study, the results residues from Glnlz3toC-terminal Ala':'9 contains a site iments provided us with supportingevidence for the putative molecules involved in binding to the cognate DNAs (12, 13). However, view that the phosphorylation event on OmpR it is also clear that theDNA binding abilityof the C-terminal induces anoligomeric interaction between them in vitro. The domain is modulated through phosphorylation of the N-ter- OmpR oligomers detectedoncross-linking with DMSon minal domain, i.e. the phosphorylation of OmpR results in SDS-polyacrylamide gels comprised the dimer andtrimer. ATP

A+ -At -L+ "

A L L -A+ -L+ - L+ -t-t-t

Gene Activation and

10780

Osmoregulation in E. coli

However, at present, the stoichiometry of the OmpR subunit in an oligomer is notyet certain, because the o m ptetramer ~

2. Csonka, L. N. (1989) Microbiol. Reu. 5 3 , 121-147 3. Stock, J. B., Ninfa, A. J., and Stock, A. M. (1989) Microbial. Reu. 53,450-490 was detected in Some experiments. In Order to 4, Mizuno, T., and Mizushima, S. (1990) Mol, Microbial. 4, 1077this, experimental conditions for more efficient in vitro cross1082 linking should be explored. In anyevent, it should be empha5. Aiba, H., Nakasai, F., Mizushima, S., and Mizuno, T. (1989) J . Biochem. (Tokyo) 106,5-7 sized here that the mutant OmpR proteins characterized in this study were defective in this particular in vitro OmpR 6. Aiba, H., and Mizuno, T. (1990) FEBS Lett. 2 6 1 , 19-22 7. Nara, F., Matsuyama, S., Mizuno, T., and Mizushima, S. (1986) oligomerization induced by OmpR phosphorylation. Recalling Mol. & Gen. Genet. 202,194-199 that cells containing thesemutant proteins exhibitthe OmpF8. ~ i T., K ~ ~ M.,~~J ~~Y.-L,, , ,~ and Mizushirna, ~ , S. (1988) J , Biol. Chem. 263,1008-1012 OmpC- phenotype, we propose that the phosphorylation9. Aiba, H., Nakasai, F., Mizushima, S., and Mizuno, T. (1989) J. induced OmpR oligomerization observed in vitro may be physBiol. Chem. 264,14090-14094 iologically relevant and may be essentially involved in the mechanism underlying the efficient recognition of the o m p ~ 10. Kanamam, K., A h , H., Mizushima, S., and Mizuno, T. (1989) J. Biol. Chem. 264,21633-21637 and OmPC promoter DNAs in Furthermore* two amino 11. Kanamaru, K., Aiba, H.,and Mizuno, T. (1990) J. Biochem. acid residues, Glug6 and Argil5, located in the N-terminal (Tokyo) 1 0 8 , 483-487 domain of OmpR were suggested to be involved in thepresum- 12. Tate, S., Kato, M., Nishimura, Y., Arata, Y., and Mizuno, T. (1988) FEBS Lett. 2 4 2 , 27-30 able OmpR-OmpR interaction. In order to further substanti13. Kate, M., A h , H., Tat'% s.7 Nishimura, y.3 and Mizuno, T. ate our proposal, extensive experimental approaches should (1989) FEBS Lett. 2 4 9 , 168-172 be carried out since, in this study, we only applied in vitro 14. Igo, M. M,, and Silhavy, T, J. (1988) J , Bacterial. 170, 5971cross-linking analysis. Experiments along theselines arecur5973 15. Aiba, H., Mizuno, T., and Mizushima, S. (1989) J. Biol. Chem. rently under way in our laboratory. The envZll mutation is known to confer on cells a pleio264,8563-8567 tropic phenotype, namely ~ ~ phoA-, 1 - through a yet 16. Forst, s.,Delgado, J., and InoUYe, M. (1989) i h c . N d . A d . Sci. U. S. A . 8 6 , 6052-6056 unknown mechanism (26)* it be briefly men- 17. Garrett, S., Taylor, R. K., Silhavy, T. J., and Bermann, M. L. tioned that the mutant ompR alleles characterized in this (1985) J. Bacteriol. 1 6 2 , 840-844 study were originally isolated, through sophisticated proce- 18. Mizuno, T., and Mizushima, S. (1987) J. Biochem. (Tokyo) 101, 387-396 dures, as codominant pseudo-suppressors for the Mal- phenotypebut not for the and phoA- phenotypes exhib- 19. Matsuyama, s.,Mizuno, T., and Mizushima, s. (1986) J. Backriol. 1 6 8 , 1309-1314 by the mutation* it has been proposed 20. Sanger, F., Nicklen, s., and Coulson, A.R. (1977) Proc. Natl. that the envZll allele exhibits the pleiotropic effect via the Acad. Sci. U. S. A . 74,5463-5467 ompR function (28), at present it is difficult to explain the 21. Jo, Y.-L., Nara, F., Ichihara, S., Mizuno, T., and Mizushima, S. (1986) J. Biol. Chem. 2 6 1 , 15252-15256 mechanism by which the mutant ompR96A and ompRl15S H., Staehelin, T., and Gordon, J. (1979) Proc. Natl.A c d . alleles can partially and codominantly suppress the defect, 22. Towbin, Sci. U. S. A . 76, 4350-4354 Mal-, caused by the envZl1 mutation. Further genetic study 23. ~ ~ U, ~ K,~(1979) ~ N~~~~~ l i 227, , 680-685 on these particular mutant ompR alleles will facilitate clari- 24. Choi, D.-S., Yamada, H., Mizuno, T., and Mizushima, S. (1987) J . Biochem. (Tokyo) 102,975-983 fication of the complex mechanism underlying the pleiotropic effect of the envZ11 mutation. Experiments along these lines 25. Hall, M. N.9 and Silhavy, T. J. (1981) J . MOL Bid. 1469 23-47 26. Wandersman, C., Moreno, F., and Schwartz, M. (1980) J. Bucteare currently under way in our laboratory. riol. 143, 1374-1383 27. Nara, F., Mizuno, T., and Mizushima, S. (1986) Mol. & Gen. REFERENCES Genet. 205,51-55 1. Hall, M. N., and Silhavy, T. J. (1981) Annu. Reu. Genet. 15,9128. Slauch, J. M., Garrett, S., Jackson, D. E., and Silhavy, T. J. 142 (1988) J . Bacteriol. 170, 439-441

omp~-

ornp~-