Proc. Nati. Acad. Sci. USAVol. 83, pp. 8278-8282, November 1986Genetics

A gene essential for Agrobacterium virulence is homologous to afamily of positive regulatory loci

(Agrobacterium virG sequence/regulation/protein homology)

STEPHEN C. WINANS*, PAUL R. EBERT*t, SCOTr E. STACHEL*I% MILTON P. GORDON§,AND EUGENE W. NESTER*Departments of *Microbiology and Immunology and §Biochemistry, University of Washington, Seattle, WA 98195

Communicated by Luis Sequeira, July 9, 1986

ABSTRACT The vir region of Agrobacterium tumefaciensspans at least six transcriptional loci required for crown galltumorigenesis. The transcriptional induction oftwo of these virloci in response to cocultivation with tobacco suspension cellswas measured by using bacteria containing mutations in eachof the six vir loci located on the Ti plasmid. Induction of thesevir genes occurred only in bacteria that had functional copiesof virA and virG. The nucleic acid sequence of a 1.25-kilobaseclone encompassing virG contains one open reading framecapable of coding for a protein of about 30,000 daltons. Theamino acid sequence of the predicted virG product is homolo-gous to that of eight bacterial proteins, including that of theompR gene ofEscherichia coli. Most, although not all, of theseproteins, like VirG, are positive regulatory elements.

Several species of bacteria that form parasitic or symbioticassociations with higher plants have been shown to inducethe transcription of genes required for these associationswhen cocultivated with plant cells or tissues. Agrobacteriumtumefaciens responds to cocultivation with plant suspensioncells by inducing the transcription of five of the six virtranscriptional units carried by the Ti plasmid. virB, virC,virD, virE, and virG are induced between 10- and 300-fold bycocultivation, whereas transcription of virA is unaffected (1).Each of these genes is required for a fully virulent phenotypeon all plant hosts, although a severely attenuated virulencehas been observed with strains having mutations in virC orvirE (2). The pinF gene, which is not required for virulence,is also induced by cocultivation (1). The products of the virgenes are thought to mediate the transfer of a segment ofplasmid DNA (the transferred DNA) from the bacterium intothe nucleus of the infected plant cell, where it becomescovalently integrated into genomic DNA (ref. 3; for review,see ref. 4). vir gene transcription can also be induced byacetosyringone and a-hydroxysyringone, two related phe-nolic compounds released by wounded, metabolically activeplant cells (5). Transcription is also inducible by the additionof a mixture of plant-derived phenolic compounds (6). Sim-ilarly, a Rhizobium meliloti gene required for nodulation isinduced by a low molecular weight, heat-stable factor ob-tained from alfalfa root exudates (7).At present, little is known about the mechanism by which

these environmental signals affect gene expression. A muta-tion on the Agrobacterium chromosome has been shown toenhance the basal rate of transcription of virC and virD (8).In the present study, transcriptional fusions between virpromoters and the Escherichia coli lac operon were used todetermine whether any of the six plasmid-encoded vir lociplay a role in the regulation of these genes. Two vir geneswere indeed required for induction, and nucleotide sequenc-

ing ofone ofthese genes revealed protein sequence homologywith a number of bacterial proteins, most of which arepositive regulatory elements.

MATERIALS AND METHODS

Materials. Restriction endonucleases were purchased fromBethesda Research Laboratories or New England Biolabsand used according to the supplier's recommendations. DNApolymerase I, Klenow fragment, was purchased fromBoehringer Mannheim or Bethesda Research Laboratories;S1 nuclease was from New England Biolabs, and E. coliexonuclease III and T4 ligase were from Bethesda ResearchLaboratories. The M13 sequencing primer, unlabeled nucle-otides, and dideoxynucleotides were obtained from P-LBiochemicals. 35S-labeled dATP was purchased from NewEngland Nuclear.

Strains and Plasmids. Agrobacterium strain A136 is aplasmid-free derivative of strain C58. A348 is a derivative ofA136 containing the octopine-catabolizing plasmid pTia6NC(9). pRK404 was obtained from D. Helinski (10). Plasmidswere mutagenized with Tn5 as described (11). Plasmid DNAwas isolated by the method of Birnboim and Doly (12).pSM243cd and pSM358cd were constructed by digestingpVK219 virB243: :Tn3HoHol and pVK225 virE358::Tn3HoHol (13), respectively, with Sal I, ligating thesefragments to each other, and using them to transform an E.coli strain to resistance to carbenicillin and kanamycin. Sal Ihas several recognition sites within the Ti plasmid DNA andcuts at the junctions between the Ti plasmid DNA and thevector pVK102 but does not cut either within the vector orwithin Tn3HoHol. Therefore, pSM243cd and pSM358cdcontain the vector pVK102 and the Sal I fragment containingthe Tn3HoHol insertion but no other Ti plasmid DNA.

Cocultivation with Plant Cells. Bacteria were cultured toOD6w- 1.5 in MG/L broth (14) containing 20 ,ug ofkanamycin per ml, resuspended in an equal volume of MSplant cell medium (15) containing 0.2 ,g of 2,4-dichlorophe-noxyacetic acid per ml, and diluted 20:1 into 6.5-ml NTIIBtobacco suspension cultures that had been subcultured 3 dayspreviously. Incubations were carried out in 20-ml test tubeson a reciprocal shaker (160 rpm) at 28-30°C. At various timeintervals, a 1.5-ml sample was removed from each cultureand frozen at -70°C. After collection of the last samples, allsamples were thawed, the plant cells were allowed to settle,and the supernatant containing the bacteria was removed andassayed for ,B3galactosidase activity as described (16).

Abbreviations: kb, kilobase(s); ORF, open reading frame(s).tPresent address: Institute of Biological Chemistry, WashingtonState University, Pullman, WA 99164.tPresent address: Institute of Cancer Research and Howard HughesMedical Institute, College of Physicians and Surgeons, ColumbiaUniversity, New York, NY 10032.

8278

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

July

9, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 8279

50

0 . 70 3. 70C',~~ ~ ~ ~~~ie;hr

CD

30N~~~~~~20(

*1

oz

0 3.5 7 3g-~~ ~ ~ ~~~imh

h

0 3.5 7

Nucleotide Sequencing. DNA sequencing was performed bythe chain-termination method (17) using 3"S-labeled nucleo-tides (18). Progressive deletions of M13 clones were createdby digestion with exonuclease III and Si nuclease (19). Bothstrands were sequenced over their entire lengths.

RESULTSRole of vir Loci in the Regulation of virB and virE. We have

isolated a large collection of transposon insertion derivativesof several cosmid clones containing Ti plasmid DNA usingthe transposon Tn3HoHol (1, 13). Since Tn3HoHol generatestranscriptional and translational fusions between genes intowhich it transposes and the lac operon (20), many of theresulting cosmids contained fusions between vir promotersand lac and respond to cocultivation with plant suspensioncultures by inducing expression of intracellular f3-galac-tosidase. Two of these, pVK219 virB243::Tn3HoHol andpVK225 virE358::Tn3HoHol, contain a virB-lacZ and avirE-lacZ fusion, respectively. The expression of f-galac-tosidase in the wild-type A. tumefaciens strain (A348) con-taining either plasmid is induced 50- and 100-fold, respec-tively, by cocultivation (13). From these plasmids we con-structed pSM243cd and pSM358cd. These plasmids containthe virB and virE promoters, respectively (21), fused to lacZbut contain only 3.7 and 2.1 kilobases (kb), respectively, ofTi plasmid DNA. pSM243cd and pSM358cd were introducedinto strain A348. P-Galactosidase activity of both transcon-jugants was strongly induced by cocultivation with plant cells(Fig. la). The kinetics of induction of the two fusions weresimilar, full induction occurring within the first 3.5 hr ofcocultivation. The virE::lacZ fusion resulted in about 20-foldhigher 3-galactosidase activity than the virB::lacZ fusion,possibly due to higher levels of transcription, greater stabilityof the virE: :lacZ fusion protein, or a combination of the two.To test whether induction depended upon a gene located on

the Ti plasmid, pSM243cd and pSM358cd were introducedinto the Ti plasmidless strain A136. Induction was notdetectable for either strain (Fig. lb), indicating that regulationof virB and virE depends upon one or more genes carried bythe Ti plasmid. pSM243cd and pSM358cd were also intro-duced into derivatives of A348 containing transposon muta-tions in one of six vir loci. As shown in Fig. 1 c-h, mutationsin either the virA or virG transcriptional units abolishedinduction of,-galactosidase, whereas mutations in virB,virC, virD, or virE did not affect induction. Two otherinsertion mutations in virB also had no effect on induction

FIG. 1. Transcriptional induction ofa virB-lacZC and a virE-lacZ fusion by cocultivation with tobac-

co suspension cells. Agrobacterium strains contain-- ming pSM243cd (o) or pSM358cd (A) were coculti-

o X vated with NTIIB tobacco suspension cultures. Inr * I a-h, pSM243cd activity has been plotted against the

cx w left axis (0-50 units activity), whereas pSM358cdactivity has been plotted against the right axis

t? (0-1000 units activity). (a) Strain A348, whicht contains Ti plasmid pTiA6 (wild type). (b) A136,

500 which lacks the Ti plasmid. (c) A348 virA226MX.(d) A348 virBl3. (e) A348 virC364MX. (W) A348virD29. (g) A348 virE34. (h) A348 virG19 (3; 13). Inparallel experiments, bacteria and plant cells wereincubated separately, and in those experiments,

0 f-galactosidase activity of each culture did notincrease detectably above initial levels even afterincubation up to 7 hr.

(data not shown). These results indicate that virA and virG actin a positive manner to regulate the transcription ofboth virBand virE. Experiments are necessary to determine whetherthese genes play a role in the regulation of virC, virD, virG,and pinF.

10 20 30 40 50 60 70AGATCTGCCT CGCGGCGGAC GCACGACGCC G tGCAAG CATAGGCGAT CTCCTAAATC AATAGTAGCT

80 90 100 110 120 130 140GTMCCTCGA AGCGTTTCAC TTGTMCAAC GATTGAGMT TTTTGTCATA AAATTGMAAT ACTTGGTTCG

150 160 170 180 190CATTTTTGTC ATCCGGGCAG CCGCMTTCT GACGAACTGC CCATTTAGCT GGAG ATG ATT GTA

MET lie Val

209 224 239 254CAT CCT TCA CGT GAA MT TTC TCA AGC GCT GTG AAC AAG GGT TCA GAT TTT AGAHis Pro Ser Arg Glu Asn Phe Ser Ser Ala Val Asn Lys Gly Ser Asp Phe Arg

269 284 299TTG AAA GGT GAG CCG TTG AM CAC GTT CTT CTT GTC GAT GAC GAC GTC GCT ATGLou Lys Gly Glu Pro Lou Lys His Val Lou Lou Val Asp Asp Asp Vol Ala MET314 329 344 359CGG CAT CTT ATT ATT GM TAC CTT ACG ATC CAC GCC TTC AAA GTG ACC GCG GTAArg His Leu lle Ile Glu Tyr Lou Thr Ile His Ala Ph. Lys Val Thr Ala Val

374 389 404 419GCC GAC AGC ACC CAG TTC ACA AGA GTA CTC TCT TCC GCG ACG GTC GAT GTC GTGAla Asp Ser Thr Gin Phe Thr Arg Val Lou Ser Ser Ala Thr Val Asp Vol Vol

434 449 464GTT GTT GAT CTA MT TTA GTT CGT GMA GAT GGG CTC GAG ATC GTT CGT MT CTGVal Val Asp Lou Asn Lou Val Arg Glu Asp Gly Lou Glu Ile Val Arg Asn Lou

479 494 509 524GCG GCA MG TCT GAT ATT CCA ATC ATA ATT ATC AGT GGC GAC CGC CTT GAG GAGAla Ala Lys Ser Asp Ile Pro lie Ile lie lie Ser Gly Asp Arg Lou Glu Glu

539 554 569ACG GAT AM GTT GTT GCA CTC GAG CTA GGA GCA AGT GAT TTT ATC GCT MG CCGThr Asp Lys Vol Val Ala Lou Glu Lou Gly Ala Ser Asp Phe Ile Ala Lys Pro

584 599 614 629TTC AGT ATC ACA GAG TTT CTA GCA CGC ATT CGG GTT GCC TTG CGC GTG CGC CCCPhe Ser Ile Arg Glu Phe Lou Ala Arg Ile Arg Val Ala Lou Arg Vol Arg Pro

644 659 674 689AAC GTT GTC CGC TCC AM GAC CGA CGG TCT TTT TGT TTT ACT GAC TGG ACA CTTAsn Val Val Arg Ser Lys Asp Arg Arg Ser Phe Cys Phe Thr Asp Trp Thr Lou

704 719 734AAT CTC AGG CAA CGT CGC TTG ATG TCC GAA GCT GGC GGT GAG GTG AM CTT ACGAsn Lou Arg GIn Arg Arg Lou MET Ser Glu Ala Gly Gly Glu Val Lys Lou Thr

749 764 779 794GCA GGT GAG TTC MAT CTT CTC CTC GCG TTT TTA GAG AM CCC CGC GAC GTT CTAAla Gly Glu Phe Asn Lou Lou Lou Ala Phe Lou Glu Lys Pro Arg Asp Val Lou

809 824 839TCG CGC GAG CAA CTT CTC ATT GCC AGT CGA GTA CGC GAC GAG GAG GTT TAT GACSer Arg Glu Gin Lou Lou lie Ala Ser Arg Vol Arg Asp Glu Glu Val Tyr Asp854 869 884 899AGG AGT ATA GAT GTT CTC ATT TTG AGG CTG CGC CGC A CTT GAG GCG GAT CCGArg Ser lie Asp Val Lou le Lou Arg Lou Arg Arg Lys Lou Glu Ala Asp Pro

914 929 944 959TCA AGC CCT CM CTG ATA AM AGA GCA AGA GGT GCC GGT TAT TTC TTT GACGCGSer Sor Pro Gin Lou lie Lys Thr Ala Arg Gly Ala Gly Tyr Ph. Phe Asp Ala

974 989 1008 1018GAC GTG CAG GTT TCG CAC GGG GGG ACG ATG GCA GCC TGA GCCCTTGCA TTTGCCTCTTAsp Vol Gin Vol Ser His Gly Gly Thr MET Ala Ala

1028 1038 l048 le58 l068 1078 1088AATTATCTGG CTCAAAGGGT GACTGACGAG TMGCGATGT GCCCATCACA CTGACCACCA AGACGGGT

1098 1108 1118 1128 1138 1148 1158TCGTTUTGM GATCACCCTT CAGCCAAG GCCAGTGTGC GTTGATCGCG GCCACACTGC AMTTGACAT1168 1178 1188 1198 1208 1218 1228

ACTGTCGATA AAGGCGTT GGAGCAGAAC CCATTCGTCA GGATATGTCC TCGTGATCTG GCGGTGCTCC1238 1248

GTCTGGTGCT CGAICCACC TGCAG

FIG. 2. Nucleic acid and predicted protein sequence of the virG gene.

I

Dow

nloa

ded

by g

uest

on

July

9, 2

020

8280 Genetics: Winans et al.

Nucleotide Sequence Analysis of the virG Locus. The virGcomplementation group is represented by three insertions ofTn3HoHol that fall within a 0.3-kb region. A 1.25-kb DNAfragment of the region containing the three virG insertions,flanked on the left by a Bgi II site and on the right by a PstI site, was cloned into plasmid vector pRK404 digested withBamHI and Pst I. Since pRK404 expresses resistance only totetracycline, which is expressed poorly in Agrobacterium, akanamycin-resistant derivative, pSW103, was constructed,by transposition of Tn5 into vector DNA. pSW103 comple-mented all three virG mutants for virulence (13) and thereforeencodes a fully functional virG gene product.The nucleotide sequence of this Bgl II-Pst I fragment was

determined (Fig. 2). Two open reading frames (ORF) werefound within this interval, a 267-codon ORF reading from leftto right (Fig. 2) and a 210-codon ORF reading from right toleft. The initiation and termination codon of the smaller ORFcorrespond to the sequences CAT (bases 815-817) and TTA(bases 185-187), respectively (Fig. 2). The two ORF there-fore overlap over much of their lengths. We believe that thevirG transcriptional unit consists of a single gene encoded bythe larger ORF for two reasons. (i) Of the three insertions ofTn3HoHol isolated in virG, only the two that are transcribedfrom left to right are transcribed at detectable levels and areinduced by cocultivation (1, 20). In addition, the size of eachof the induced virG: :lacZ fusion proteins corresponds to thepredicted size if translation initiates at the initiation ATGcodon of the 267-codon ORF (20). (ii) We have localized atranscription initiation site upstream of the 267-codon ORFby S1 nuclease mapping (21). The significance of the smallerORF to the biology of tumorigenesis is unknown.We do not see a sequence homologous to the E. coli

ribosome binding site even though all other vir genes whosesequences have been determined do have such homology(21). Similarly, the transcription initiation site, the guanine atposition 148 (21), is not preceded by a sequence homologousto the -10 sequence of E. coli despite the fact that other virgene promoters do have homologous regions. These findingssuggest that the VirG protein may be expressed at relativelylow levels. The protein predicted by the 267-codon ORF has

VI R GOM P RS FR AP HO SC H E SS PO 0 ANT R CS P0 0 FCH E Y

VI R GOM P RS FR AP HOBC H E BS PO 0 AN TR CS P0 0 FCH E Y

VI R GO MP RS FR AP HO BC H E BS PO 0 ANT R CS P0 0 FC HE Y

VI R GOM P RS FR AP HO SVI R GOM P RS FR AP HOB

M I V H P S R E N F S S A V N K C S D F R

E AE at

L ~

G 1_~- 5sM.

IA_

EEysLG T V E

Proc. Natl. Acad. Sci. USA 83 (1986)

a molecular mass of 30,083 daltons and a net charge of +1.The large number of charged groups along the molecule andthe lack of any significant hydrophobic domains suggest thatthe virG product may be a soluble protein. There is noapparent DNA binding motif common to most DNA bindingproteins (22).Computer-Aided Search for Proteins Homologous to VirG.

The data bank of the National Biomedical Research Foun-dation (3061 sequences, August 1985 revision) was searchedfor sequences homologous to the predicted VirG protein withthe FASTP program (23). The 267-codon ORF was homolo-gous to the proteins coded by the ompR locus of E. coli andthe spoOA gene ofBacillus subtilis. (For a description ofthesegenes, see Discussion.) It has been reported that ompR ishomologous to the products of the sfrA and phoB genes (ref.24; K. Makino, H. Shinagawa, M. Amemura, and A. Nakata,personal communication) of E. coli, the cheB and cheYproducts ofE. coli (25) and Salmonella typhimurium (26), thentrC gene of Klebsiella pneumoniae (27), and the spoOFgene of B. subtilis (28). We therefore compared the aminoacid sequence of VirG with that of the other eight proteinsusing the FASTP algorithm.

Fig. 3 shows the amino acid sequences of the products ofthe virG, ompR, sfrA, phoB, che Y, and spoOF genes and thefirst 130 amino acids of the cheB, spoOA, and ntrC products.To optimize alignments between these eight proteins, smallgaps indicated by dashes were introduced by the FASTPalgorithm and have been modified slightly by hand. VirG ishomologous over its entire length to OmpR, SfrA, and PhoB,whereas homology with the other five proteins extended onlyover the amino-terminal half of VirG (Table 1). CheY andSpoOF proteins are considerably shorter than VirG and showhomology over their entire length to amino acids 25-160 ofVirG. In contrast, CheB, SpoOA, and NtrC are longer thanVirG, but there is no evident amino acid homology betweenVirG and the other three beyond the sequences shown. Theregion common to all nine proteins has been designateddomain A, whereas the carboxyl-terminal regions of VirG,OmpR, SfrA, and PhoB have been designated domain B (Fig.4).

L KG EPM ENYN N

M a T P

V S KM EK IKVAR GIV N aI

VMAD K Ed-

DPI ~~~~~~~~~~~~~~~~~~~~~~IRPmD P SEW

RPIBY IFX

II

K _A PA

10TISS

-E - P_P _Y

-. P.

ILSRR

A_

*-

D-D

V _ _ ,P G_IP G Ils_P G 1 - I

PpI L E IP G M IP G M IP am I

ARRK R T R aR L]---R - L KK ;H - - -KR IH:VR IIAM S

- -P;3.F,; Rl , ED V; HGGTPME AV-

_zp v_-1 P ^ - - -~~

V IE

LMM_ L A i N E V I A E K V R T A . .. .1-E~~ ~~~_Q G ~T . .. .-E-D R ERM s H Y 0 E O Q Q P R N . .. .

V~

E K I F E K L G V

_ RI AP.I I . .K .. : . ; *P :; ! IN L_i5 5 E T ~~~~~~~~~~KK TG3

FDn V S H G G T V A A_V XV ~~~~~~~~GS K A

- s E _R sD~~~RFIG. 3. Alignment of VirG with eight homologous proteins. Optimal alignments were found with the FASTP algorithm. Amino acid residues

shown in black boxes represent amino acids identical or similar between VirG and its homologues. Amino acids 1-130 ofCheB, SpoOA, and NtrCare shown.

Dow

nloa

ded

by g

uest

on

July

9, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 8281

Table 1. Amino acid sequence homologies between the VirG,OmpR, SfrA, PhoB, CheB, SpoOA, NtrC, SpoOF, andCheY proteins

Proteins % Similaritycompared Amino acid positions identity (SD)

Domain AVirG vs. OmpR 25-150 vs. 1-126 45 26VirG vs. SfrA 25-150 vs. 1-124 30 21VirG vs. PhoB 25-150 vs. 1-126 28 20VirG vs. CheB 25-150 vs. 1-128 32 15VirG vs. SpoOA 25-150 vs. 1-128 27 11VirG vs. NtrC 25-150 vs. 1-124 26 12VirG vs. CheY 25-150 vs. 1-129 25 14VirG vs. SpoOF 25-150 vs. 1-124 25 14OmpR vs. SfrA 1-126 vs. 1-124 41 28OmpR vs. PhoB 1-126 vs. 1-126 43 29OmpR vs. CheB 1-126 vs. 1-128 30 9OmpR vs. SpoOA 1-126 vs. 1-128 35 14OmpR vs. NtrC 1-126 vs. 1-124 29 18OmpR vs. CheY 1-126 vs. 1-129 26 15OmpR vs. SpoOF 1-126 vs. 1-124 32 14SfrA vs. PhoB 1-124 vs. 1-126 35 23SpoOA vs. SpoOF 1-128 vs. 1-124 32 14CheB vs. CheY 1-128 vs. 1-129 28 6

Domain BVirG vs. OmpR 151-269 vs. 127-240 40 20VirG vs. SfrA 151-269 vs. 125-239 35 27VirG vs. PhoB 151-269 vs. 127-239 36 13OmpR vs. SfrA 127-240 vs. 125-239 31 17OmpR vs. PhoB 127-240 vs. 127-230 34 14SfrA vs. PhoB 125-239 vs. 127-230 31 18

The amino acid sequences of the proteins were aligned using theRDF algorithm of Lipman and Pearson (23). Similarity scores foreach pair ofproteins are expressed as a standard deviation (SD) fromthe mean of the aligned scores obtained when the second of the twosequences was randomly permuted 50 times. Values of 3-6 SDsignify that the two proteins possibly share a common ancestry, 6-10SD, a probable common ancestry, and >10 SD, a very probablecommon ancestry (see ref. 23).

DISCUSSIONWe have examined the role of six vir loci, located on theAgrobacterium plasmid pTiA6, on induction of virB and virEduring cocultivation with plant suspension cells. Inductiondepended on functional copies of virA and virG. An analysisof the nucleotide sequence of virG revealed one ORF that iscapable of coding for a protein of about 30,000 daltons. Theprotein sequence predicted by the virG nucleotide sequenceis homologous to a class of proteins that may have similarroles in diverse systems.

VirG

OmpR

SfrA

PhoB

CheB

SpoOA

NtrC

CheY

SpoOF

A B

- 4 $-I4I

Il

Most of the members of this gene family code for positiveregulatory elements. In the present study, we have shownthat VirG is required for the positive regulation ofat least twovir loci coded by the Ti plasmid ofAgrobacterium. OmpR isrequired for expression of outer membrane porins coded byompC and ompF (29) and for genes coding for microcin 17 andcolicin E2 (30). SfrA (sex factor regulation) is required fortranscription of the traJ gene of the F plasmid as well as anumber of membrane proteins (31). PhoB is required toregulate the phosphatase regulon includingphoA (32). SpoOAand SpoOF are required for the expression of genes requiredfor sporulation (28, 33). NtrC is required for expression ofnitrogen-regulated genes, including nifAL and ginA (27). Incontrast, CheB and CheY do not appear to regulate tran-scription but, rather, are required for bacterial chemotaxis.CheB is a methylesterase that removes methyl groups frommembrane-bound chemoreceptors (34). CheY is believed tointeract directly with the CheR methyltransferase and withthe flagellar basal structure to control cell motility (26). Thefinding that the majority of the proteins homologous to VirGare positive regulatory elements lends support to our geneticstudies indicating that VirG has a similar function andsuggests that VirG may act by mechanisms similar to thoseof one or more of its homologues.

All of these proteins are components of information-transducing pathways, in that each provides a mechanism bywhich bacteria can obtain information from their environ-ment and carry out appropriate responses. For example,VirG allows information about the presence of plant metab-olites to result in the induction of vir genes. Similarly, OmpRtransmits information about the osmotic pressure of theenvironment to the ompC and ompF genes. SpoOA, SpoOF,PhoB, and NtrC cause information about the cell's nutritionalstatus to result in expression of appropriate genes. CheB andCheY enable the cell to respond to temporal changes in theconcentration of chemoattractants and repellants with ap-propriate alterations in cell motility.On the basis of the sequence homologies found between

VirG and the other eight proteins, we propose that VirG maycontain two domains [this model is a modified version of oneoriginally proposed by Drummond et al. (27)]. Domain A isshared by all nine proteins, whereas domain B is shared byonly four of them (Fig. 4). There is evidence that domain Aof at least four of these proteins possesses some degree offunctional autonomy. Two of them, CheY and SpoOF, whichmust be capable of some function, fall almost entirely withinthis domain (Fig. 4). The other two, CheB and NtrC, areconsiderably larger but, in both cases, there is evidence thatthe portions falling within domain A do play a semiautono-mous role. CheB is susceptible in vivo to proteolytic cleavage

FIG. 4. Domain structure ofVirG and eight homologues. Theregions falling within domains Aand B and their degrees of simi-larity are shown in Table 1.

Genetics: Winans et al.

I

Dow

nloa

ded

by g

uest

on

July

9, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986)

at the carboxyl terminus of domain A, and molecules lackingthe A domain have a 15-fold higher methylesterase activitythan the intact protein (34). This finding indicates that domainA is totally dispensable for methylesterase activity and,furthermore, that it may in some way modulate the activityof that portion of the enzyme containing the catalytic site.Similarly, there is genetic evidence (27) that domain A ofNtrC is responsible for interconverting this protein betweenits activator and its repressor forms. It, too, may therefore actto modulate an activity lying elsewhere in the NtrC protein.By analogy, we propose that domain B of VirG (and itshomologues) may contain the sites required to promotetranscription and that domain A may modulate the activity ofdomain B. Our knowledge of the nucleotide sequence of virGwill allow us to make the precise genetic alterations requiredto test this model.Many, if not all, of the proteins homologous to VirG

require the products of other genes in order to function. In thepresent study we have shown that VirG regulates vir loci onlyin strains that are proficient in virA. Similarly, OmpRfunction depends upon the envZ product (35), SfrA uponcpxA (31), PhoB upon phoR (32), SpoOA and SpoOF uponspoOB and spoOE (28, 33), and NtrC upon ntrA and ntrB (36).The nucleic acid sequences of virA (B. Leroux, S. Zeigler,and E.W.N., unpublished data), envZ (35), cpxA (37), phoR(H. Shinagawa, personal communication), and ntrB (38) havebeen determined, and our analysis of the proteins encoded bythese five loci indicates that all are homologous. The virAtranscriptional unit encodes a single protein, which resem-bles EnvZ and various transmembrane chemoreceptor pro-teins (39) in that it contains two hydrophobic regions, one atthe amino terminus and the other about 250 amino acidstoward the carboxyl terminus. It has been proposed (35, 39)that approximately one-third ofEnvZ and of the chemorecep-tor proteins (near the amino terminus) is periplasmicallylocalized and that the remainder of these proteins is locatedin the cytoplasm. Hall and Silhavy (29) have proposed thatEnvZ is able to perceive osmotic pressure and to use thisinformation to modulate the ability of OmpR to promote

0o o WOUND-INDUCED0(

0 PLANT PHENOLICS

0

PERIPLASM

CYTOPLASM

VirG °00

INACTIVE

AA

ACTIVE

virBv r B

v i r E

FIG. 5. Model describing how VirA and VirG may cause thepresence of wound-induced plant metabolites to result in the expres-sion of plasmid-encoded vir genes.

transcription. We believe that VirA, like EnvZ, may span theinner membrane and, by analogy to EnvZ, that it may directlysense the presence of plant-derived phenolic compounds(Fig. 5). After receiving this signal, VirA could transformVirG from an inactive form to an active form capable ofpromoting the transcription of genes under its control. Asimilar model has been proposed (40) to describe the controlof nitrogen-regulated genes by NtrB and NtrC.

We thank H. Shinagawa for providing unpublished data and T.Nixon, C. Ronson, and F. Ausubel for sharing unpublished insightsabout the homology between ntrB and envZ. This work was sup-ported by National Institutes of Health Grant 5 ROI GM32618-14 andNational Science Foundation Grant PCM-8315826. P.R.E. wassupported by Molecular and Cellular Training Grant 5T32GM07270-12. S.C.W. was supported by Damon Runyon-Walter WinchellCancer Fund Fellowship DRG-800.

1. Stachel, S. E., Nester, E. W. & Zambryski, P. C. (1986) Proc. Natl.Acad. Sci. USA 83, 379-383.

2. Klee, H. J., White, F. F., Iyer, V. N., Gordon, M. P. & Nester, E. W.(1983) J. Bacteriol. 153, 878-883.

3. Koukolikova-Nicola, Z., Shillito, R. D., Hohn, B., Wang, K., VanMontagu, M. & Zambryski, P. (1985) Nature (London) 313, 191-196.

4. Nester, E. W., Gordon, M. P., Amasino, R. M. & Yanofsky, M. F.(1984) Annu. Rev. Plant Physiol. 35, 387-413.

5. Stachel, S. E., Messens, E., Van Montagu, M. & Zambryski, P. (1985)Nature (London) 318, 624-629.

6. Bolton, G. W., Nester, E. W. & Gordon, M. P. (1986) Science 232,983-985.

7. Mulligan, J. & Long, S. (1985) Proc. Nati. Acad. Sci. USA 82,6609-6613.

8. Close, T. J., Tait, R. C. & Kado, C. I. (1985) J. Bacteriol. 164, 774-781.9. Sciaky, D., Montoya, A. L. & Chilton, M.-D. (1977) Plasmid 1,

238-253.10. Ditta, G., Schmidhauser, T., Yakobson, E., Lu, P., Liang, X.-W.,

Finlay, D. R., Guiney, D. & Helinski, D. R. (1985) Plasmid 13,149-153.11. Winans, S. C. & Walker, G. C. (1985) J. Bacteriol. 161, 402-410.12. Birnboim, H. C. & Doly, J. (1979) Nucleic Acids Res. 7, 1513-1523.13. Stachel, S. E. & Nester, E. W. (1986) EMBO J. 7, 1445-1454.14. Garfinkel, D. J. & Nester, E. W. (1980) J. Bacteriol. 144, 732-743.15. Murashige, T. & Skoog, F. (1962) Physiol. Plant 15, 473-497.16. Miller, J. H. (1972) Experiments in Molecular Genetics (Cold Spring

Harbor Laboratory, Cold Spring Harbor, NY).17. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci.

USA 74, 5463-5467.18. Biggin, H. D., Gibson, T. J. & Hong, G. F. (1983) Proc. Natl. Acad.

Sci. USA 80, 3963-3965.19. Guo, L.-H. & Wu, R. (1983) Methods Enzymol. 100, 60-96.20. Stachel, S. E., An, G., Flores, C. & Nester, E. W. (1985) EMBO J. 4,

891-898.21. Das, A., Stachel, S., Ebert, P., Allenza, P., Montoya, A. & Nester, E.

(1986) Nucleic Acids Res. 14, 1355-1364.22. Pabo, C. 0. & Sauer, R. T. (1984) Annu. Rev. Biochem. 53, 293-321.23. Lipman, D. J. & Pearson, W. R. (1985) Science 227, 1435-1441.24. Drury, L. S. & Buxton, R. S. (1985) J. Biol. Chem. 260, 4236-4242.25. Mutoh, N. & Simon, M. S. (1986) J. Bacteriol. 165, 161-166.26. Stock, A., Koshland, D. E., Jr., & Stock, J. (1985) Proc. Natl. Acad.

Sci. USA 82, 7989-7993.27. Drummond, M., Whitty, P. & Wootton, J. (1986) EMBO J. 5, 441-447.28. Trach, K. A., Chapman, J. W., Piggot, P. J. & Hoch, J. A. (1985) Proc.

Natd. Acad. Sci. USA 82, 7260-7264.29. Hall, M. N. & Silhavy, T. S. (1981) Annu. Rev. Genet. 15, 91-142.30. Hernandez-Chico, C., Herrero, M., Rejas, M., San Millan, J. L. &

Moreno, F. (1982) J. Bacteriol. 152, 897-900.31. Silverman, P. M. (1985) Bioessays 2, 254-259.32. Shinagawa, H., Makino, K. & Nakata, A. (1983) J. Mol. Biol. 168,

477-488.33. Ferrari, F. A., Trach, K., LeCoq, D., Spence, J., Ferrari, E. & Hoch, J.

(1985) Proc. Natl. Acad. Sci. USA 82, 2647-2651.34. Simms, S. A., Keane, M. G. & Stock, J. (1985) J. Biol. Chem. 260,

10161-10168.35. Comeau, D. E., Ikenaka, K., Tsung, K. & Inouye, M. (1985) J.

Bacteriol. 164, 578-584.36. Dixon, R. A. (1984) J. Gen. Microbiol. 130, 2745-2755.37. Albin, R., Weber, R. & Silverman, P. M. (1986) J. Biol. Chem. 261,

4698-4705.38. MacFarlane, S. A. & Merrick, M. (1985) Nucleic Acids Res. 13,

7591-7606.39. Krikos, A., Mutoh, N., Boyd, A. & Simon, M. I. (1985) Cell 33,

615-622.40. Nixon, B. T., Ronson, C. W. & Ausubel, F. M. (1986) Proc. Natd.

Acad. Sci. USA 83, 7850-7854.

8282 Genetics: Winans et al.

Dow

nloa

ded

by g

uest

on

July

9, 2

020