JOURNAL OF BACrERIOLOGY, OCt. 1993, p. 6626-6636 Vol. 175, No. 200021-9193/93/206626-11$02.00/0Copyright C) 1993, American Society for Microbiology

The Chromosomal Response Regulatory Gene chvI ofAgrobacterium tumefaciens Complements an Escherichia coli

phoB Mutation and Is Required for VirulenceNICHOLAS J. MANTIS AND STEPHEN C. WINANS*

Section of Microbiology, Cornell University, Ithaca, New York 14853-8101

Received 7 June 1993/Accepted 6 August 1993

In an effort to identify the Agrobacterium tumefaciens phosphate regulatory gene(s), we isolated a clone froman A. tumefaciens cosmid library that restored regulated alkaline phosphatase activity to an Escherichia coliphoB mutant. The gene that complemented phoB was localized by subcloning and deletion analysis, and theDNA sequence was determined. An open reading frame, denoted chvI, was identified that encoded a predictedprotein with amino acid similarity to the family of bacterial response regulators and 35% identity to PhoB.Surprisingly, an A. tumefaciens chvI mutant showed normal induction of phosphatase activity and normal virGexpression when grown in phosphate-limiting media. However, this mutant was unable to grow in mediacontaining tryptone, peptone, or Casamino Acids and was also more sensitive than the wild type to acidicextracellular pH. This mutant was avirulent on Kalanchoe diagremontiana and was severely attenuated in virgene expression. The pH-inducible expression of virG was also abolished. Growth of the chvI mutant wasinhibited by K. diagremontiana wound sap, suggesting that avirulence may be due, in part, to the inability of thismutant to survive the plant wound environment.

Agrobacterium tumefaciens is a gram-negative phytopatho-gen that infects a wide range of dicotyledonous plants (5, 23,30, 58). Infection occurs at plant wound sites and involves thetransfer of oncogenic DNA (T-DNA) from the bacterium tothe plant cell nucleus. The T-DNA and the genes required forT-DNA transfer (vir genes) are located on a large plasmidcalled the Ti (tumor-inducing) plasmid. The vir genes arecoordinately induced in response to three environmental sig-nals: phenolic compounds secreted from plant wounds (51, 52),monosaccharides (8, 49), and acidic pH (52, 58). Induction ofthe vir genes requires two plasmid-encoded proteins, VirA andVirG (53). These proteins are members of the bacterialtwo-component regulatory system family (11, 44, 58). VirA is amembrane-bound histidine protein kinase (10, 33, 62), andVirG is a cytoplasmically located transcriptional activator (60).

In addition, several chromosomal loci participate in vir generegulation. chvE encodes a periplasmically located sugar-binding protein that interacts with VirA and stimulates virexpression in the presence of specific monosaccharides (8, 49).The ros gene product negatively regulates the virC and virDgenes but not other vir genes (14). A mutation in miaA, whichencodes a tRNA:isopentenyltransferase, decreases vir geneinduction 5- to 10-fold (25). chvD encodes a periplasmicATPase possibly involved in membrane transport (61). Disrup-tion of this gene attenuated virG induction by acidic pH andphosphate starvation, although these effects could be indirect(61). Metts et al. isolated three chromosomal loci that arenecessary for virulence and vir gene expression (40). Insertionsin any one of these loci disrupt the cellular lipopolysaccharidecomposition (40). The present study suggests that at least oneadditional chromosomal locus may be involved in vir generegulation and virulence.

Expression of the virG gene is more complex than that ofother vir genes, and at least two chromosomal regulatory

* Corresponding author.

systems regulate virG (59). virG is expressed at two tandempromoters and is responsive to three environmental stimuli:plant-released phenolics (in a VirA-VirG-dependent manner),phosphate starvation (59, 61), and acidic culture media (37, 54,55). The upstream promoter (P1) is necessary for induction byphenolic compounds (53, 59) and phosphate starvation (59),whereas the downstream promoter (P2) is induced by acidicmedia (37, 59). Induction of P1 by phosphate starvation or P2by acidic media does not require VirA-VirG or any other Tiplasmid-encoded genes, suggesting that the regulatory systemsare encoded on the chromosome. The sequence of promoterP1 is similar to those of the so-called pho boxes found in the-42 to -22 region of Escherichia coli promoters which areinduced by phosphate starvation (3, 56, 59). These promotersare coordinately regulated in E. coli by a two-componentregulatory system consisting of PhoR (histidine kinase) andPhoB (response regulator).We hypothesized that A. tumefaciens has a phosphate regu-

latory protein similar to PhoB and that it may regulate virG. Inthis paper, we identified, from an A. tumefaciens cosmidlibrary, a chromosomal gene called chvI (chromosomal viru-lence) that complemented an E. coli phoB mutation. chvI wasfound to encode a putative protein with amino acid similarityto the family of response regulatory proteins, including PhoBfrom E. coli and PhoB homologs from other bacteria. How-ever, a chvI null mutant showed normal phosphatase activityand normal virG gene induction in response to phosphatelimitation, suggesting that chvI does not function as the uniquephosphate regulator in A. tumefaciens. The chvI mutant waspleiotropic: it was unable to grow in complex media or mediabuffered to a pH below 5.5, it was avirulent on Kalanchoediagremontiana and sensitive to K diagremontiana wound sap,and it was attenuated for vir gene induction and pH-induciblevirG expression. These data suggest that chvI functions as aregulator in A. tumefaciens and is, either directly or indirectly,required for virulence.

6626

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

chvI OF A. TUMEFACIENS 6627

TABLE 1. Bacterial strains and plasmids used in this study

Strain or plasmid Relevant charaCteristics Source or reference

StrainsE. coliBW7656 lac-169 phoB23 creB510 thi 57DH5oa F- lacZAM15 A(lacZYA-argF)U169 endAl thil supE44 hsdRJ7 U.S. BiochemicalsHB101 F- leuB6 proA2 lacYl rpsL20 hsd2O supE44 recA13 18JM101 supE thi A(lac-proAB)[F' traD36 proAB lacIq lacZAM15] Laboratory collectionJM105 rpsL endoA spc A(lac-proAB)[F' traD36 proAB lacIq lacZ15] A. TorrianiJM105 phoB::Tn5 A. TorrianiSMlOXpir recA::RP4-2-Tet::Mu Kanr thi leu thr 50

A. tumefaciensA136 C58 lacking the nopaline Ti plasmid pTiC58, Rif' Nalr 47A348 A136 containing the octopine plasmid pTiA6NC 47NM103 A136 chvl::Kan This studyNM104 A348 chvI::Kan This study

PlasmidspBC6APstI phoBR in pUC18, Ampr 57pCC101 Plac::lacZ Ampr Tetr, IncP origin 13pCH114 virA Ampr Tetr, IncW origin 10pCH116 Plac::virG virB::lacZ Kanr IncP 10pGEM-7Zf Cloning vector, Ampr, ColEl origin StratagenepKNG101 Suicide vector, mobRP4 Strr, R6K origin 31pTZ18R Cloning vector, Ampr, ColEl origin U.S. BiochemicalspRK2073 Conjugal helper plasmid, Spcr, IncP origin 17pSW174 virG::lacZ SpCr, IncW origin 59pSW213 Broad-host-range cloning vector, Tetr, IncP origin 13pSW219 virB::lacZ SpCr, IncW origin 59pVJS303 pACYC184 with KINN Kanr cassette V. J. Stewart

MATERIALS AND METHODS

Reagents, enzymes, and media. Peptone, tryptone, yeastextract, Casamino Acids, and bromocresol purple were pur-chased from Difco Co. (Detroit, Mich.). Antibiotics, o-nitro-phenyl-3-D-galactopyranoside (ONPG), p-nitrophenyl phos-phate (PNPP), 2-(N-morpholino)ethanesulfonic acid (MES),N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES),and acetosyringone (AS) were purchased from Sigma Co. (St.Louis, Mo.). The chromogenic substrates 5-bromo-4-chloro-3-indolyl phosphate (X-Phos) and 5-bromo-4-chloro-3-indolyl-,B-D-galactoside (X-Gal) were purchased from Aldrich Chemical(Cleveland, Ohio) and Gold Biotechnology (St. Louis, Mo.),respectively. Restriction endonucleases and T4 DNA ligasewere purchased from Bethesda Research Laboratories (Gaith-ersburg, Md.) or New England Biolabs (Beverly, Mass.).Luria-Bertani (LB) medium (41) and AB minimal medium (9)were used for bacterial cultures. Thiamine (0.05%) was addedto AB minimal medium for E. coli cultures. AB medium wasmodified for phosphate starvation assays by replacing thephosphate buffer with HEPES (25 mM, pH 7.5). Antibioticswere used at the concentrations recommended by Sambrook etal. (46) unless noted otherwise. Sequenase was purchased fromU.S. Biochemicals (Cleveland, Ohio), and ot-35S-dATP wasfrom Amersham (Arlington Heights, Ill.). Purified phytoalex-ins were obtained from G. DiCenzo (Cornell University).

Strains and plasmids. All strains and plasmids used in thisstudy are listed in Table 1. The construction of the plasmidsused in this study is described in the text, except for that ofplasmid pNM157, which was constructed by cloning the 4.5-kbHindIll fragment from pNM139 into the vector pSW213 (13).

Transformations and bacterial matings. Plasmids were in-troduced into E. coli by transformation (46) and into A.

tumefaciens strains by electroporation (9). NM103 and NM104were prepared and electroporated as previously described (9),except that AB medium was used in place of LB and carben-icillin, when used, was added to a final concentration of 2,ug/ml. The A. tumefaciens NT-1 cosmid library was mobilizedinto JM105 phoB::Tn5 by triparental mating (17), andtransconjugants were spread on AB medium (25 mM HEPES[pH 7.5], thiamine [0.05%], tetracycline [10 ,ug/ml], kanamycin[100 ,ug/ml], K2HPO4 [0.1 mM], and X-Phos [40 ,ug/ml]).Kanamycin- and tetracycline-resistant blue colonies werepicked and streak purified. Cosmid DNA was isolated fromthese strains by a standard plasmid preparation and then usedto transform JM105 phoB::Tn5 to tetracycline resistance.Transformants were then screened for alkaline phosphatase(AP) activity as described above.Phosphate starvation assays. Bacterial cells were grown in

AB medium to an optical density at 600 nm (OD600) of 0.4 to0.6, collected by centrifugation, suspended in 0.05 volume ofwater containing 15% glycerol, and frozen at - 70°C. Forphosphate starvation experiments, cells were thawed and di-luted 1:1,000 into AB-HEPES (pH 7.5) containing K2HPO4(0.1 mM) and cultured at 30°C (A. tumefaciens) or 37°C (E.coli) on a rotary aerator for 16 to 18 h. E. coli cultures grownwith phosphate were subcultured 1:100 at 12 h into fresh mediaand then incubated an additional 4 h. AP activity was measuredessentially as described by Brickman and Beckwith (7). Sam-ples (1.5 ml) were removed, washed with 1 volume of 10 mMTris-HCl, and then suspended in 1 M Tris-HCl (pH 8.0).Depending on the anticipated AP activity, 0.1 to 0.3 ml of thesamples was removed and 1 M Tris-HCl was added to a finalvolume of 1 ml. Cells were permeabilized by adding 2 drops ofchloroform, and then 0.1 ml of PNPP (4 mg/ml) was added and

VOL. 175, 1993

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

6628 MANTIS AND WINANS

chvl

Hd St Sa PsSp Ec SaI I I I I

St

PsI

Sp

Ec

SaLM

1 kb

FIG. 1. Restriction map of pNM139 and ding the ability of these plasmids to complemeiConstruction of the plasmids is described in ttcells containing the plasmids mentioned abclow-phosphate-concentration X-Phos mediuactivity as follows: + +, strong activity; +, weaRestriction endonuclease sites EcoRI (Ec),SalI (Sa), SspI (Sp), and SstI (St) are indicat

the reaction mixtures were incubated atsamples (0.7 to 0.9 ml) were used to aflactosidase activity was measured froiNM104(pSW174) grown in AB-HEPE'mM phosphate, as described elsewhere (which contains a virG::iacZ fusion, w;NM104 to spectinomycin resistance.

Deletion mutagenesis of pNM139 and4.5-kb HindIII fragment from plasmid pboth orientations into the unique Hind]create pNM139A and pNM139B. Plasitions (see Fig. 1) were transformed into Jtransformants were screened for AP aslow-phosphate-concentration medium.was done by the dideoxy method with sinDNA (16) and Sequenase, as described(U.S. Biochemicals). Oligonucleotidesized at the Cornell biotechnology fareverse primer, which was obtained froBoth strands of the DNA were sequenc

Insertion mutagenesis of chvI. Thefragment from pNM144 containing c

pGEM-7Zf (Promega) at the SphI-HipNM153. The kanamycin resistance cascontained within a single EcoRI fragmen(Valley Stewart, Cornell University), wEthe unique EcoRI site (Fig. 1) to createBamHI fragment of pNM155 that contaallele was cloned into the suicide vectorunique ApaI-BamHI sites to create pNtained in the host SMlO0pir (50). Tcontains the sacB gene from Bacillus si

levansucrase (31), and production of levfaciens in the presence of 5% sucSMlOXpir(pNM156) was spot mated withovernight. The mating mixtures wereplated on AB minimal medium containiicin per ml. Two pools containing five i

resistant colonies of A136 or A348 were grown to mid-logI tAPi Plasmid phase and plated onto AB-kanamycin-5% filter-sterilizedattivty sucrose plates. Eighty colonies each of A136 and A348 thatHd + pNMI 31 were Kmr Sucr were purified and screened for streptomycinmm (500 ,ug/ml) resistance on LB plates. Approximately 40% ofHd the sucrose-resistant colonies were sensitive to streptomycin.

moi ++ pNMI 4 Eleven sucrose-sensitive, streptomycin-sensitive, kanamycin-

Hdresistant strains were confirmed by Southern blot analysis to

Hd .+ pNM1 4, have the desired gene disruption. For Southern blots, A.tumefaciens chromosomal DNA was purified by the hexadecyl-

Hd trimethyl ammonium bromide miniprep method (4) fromi + pNMl 4' cultures in mid-logarithmic growth. DNA (ca. 6 [Lg) was

Hd digested with Sall, size fractionated on a 0.75% agarose-TAEX ~ pNM1 41 gel, and transferred by the capillary method (46) onto a

FLASH membrane (Stratagene, La Jolla, Calif.). PlasmidHd

p

pNM142 was biotinylated with the FLASH Prime-It randomJp primer labelling kit, hybridized to the nylon membrane in

QuikHyb solution, and developed as described by the manu-leletion derivatives, show- facturer (Stratagene).nt aphoB::Tn5 mutation.he text. JMlO5phoB::Tn5 Bacterial growth curves and vir gene induction assays. Forave were plated on AB- growth curve analysis, A. tumefaciens strains were prepared asam and scored for AP described for AP assays and then diluted to an OD600 of 0.1ak activity; -, no activity. into AB minimal medium supplemented with 40 mM MES andHindIII (Hd), PstI (Ps), cultured at 30°C to an OD600 of 0.4. Cells were collected byted. centrifugation at 8,000 x g for 10 min, suspended in 1 ml of

AB minimal medium, and then diluted into 10 ml of pre-warmed AB minimal medium or LB medium (in 125-by-

t 37°C. The remaining 25-mm culture tubes) to an OD600 of 0.1 and returned to theieasure OD600 13-Ga- rotary aerator. OD was monitored with a Spectronic model 20m A348(pSW174) or spectrophotometer (Bausch and Lomb, Rochester, N.Y.). ForS with 10 mM or 0.1 vir induction assays, the inducer, AS, was added to all media to'59). Plasmid pSW174, a final concentration of 100 ,uM (59). Samples were removedas used to transform from cultures and assayed for 1-galactosidase (41).

Tumorigenesis assays. The chvI mutant was inoculated on

DNA sequencing. The K diagremontiana leaves as described by Garfinkel and NesterNM204 was cloned in (21). As controls, all test plants were independently inoculatedIII site of pTZ18R to with the wild-type parent strain (A348) and the plasmid-curedmids containing dele- derivative strain A136. A. tumefaciens cultures (3 ml) wereJMlO5phoB::Tn5, and grown in AB medium to late-log phase; 1 ml was seriallyctivity on AB-HEPES diluted, and dilutions were plated on AB plates to determineAll DNA sequencing the number of CFU per ml, and 1 ml was centrifuged, and theigle-stranded template cell pellet was applied onto the youngest full-size leaves on aby the manufacturer 10-cm stem ofK diagremontiana, which had been scraped with

primers were synthe- a wooden dowel rod to create a wound about 30 mm in length.tcility, except for the Two independent chvI mutant cultures, and the control strains,bm U.S. Biochemicals. were inoculated in duplicate at 109, 108, or 107 CFU. Infectedred. plants were stored in plastic containers out of direct light and2.1-kb SphI-HindIII in high humidity for 48 h, and then they were maintained in a

hvl was cloned into growth chamber with 50% humidity at 23°C under metal halidendIII sites to create lights as described elsewhere (10). Photographs were taken 5;sette KINN, which is weeks postinoculation.It on plasmid pVJS303 Plant sensitivity assays. A. tumefaciens strains were grownas cloned into chvI at to mid-logarithmic phase in AB minimal medium and frozen aspNM155. The ApaI- described above. Strains were thawed and then serially diluted

ins the disrupted chvI in sterile water; 0.1-ml volumes of these mixtures (ca. 106pKNG101 (31) at the CFU/ml) were added to 2.5 ml of AB top agar (AB mediumIM156 and was main- containing 0.7% agar) and overlaid on AB plates. The young-he vector pKNG101 est full-size leaves of K diagremontiana were removed at theubtilis, which encodes leaf node and sectioned into 2- to 5-mm-thick slices with a'ansucrase in A. tume- razor blade and placed on the top agar. The plates wererose is lethal (22). incubated right-side-up at 30°C for 48 h and then photo-i A136 or A348 on LB graphed.serially diluted and Nucleotide sequence accession number. The DNA sequence

ng 100 ,ug of kanamy- of the A. tumefaciens chvI gene was submitted to the GenBankindividual kanamycin- and EMBL data banks under accession number L19166.

J. BACTERIOL.

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

chvI OF A. TUMEFACIENS 6629

TABLE 2. Complementation of an E. coli phoB mutation with chvI

AP activity" at phosphate concn:Plasmid

10 mM 0.1 mM

pNM142 (chvI) 0.68 44.2pBC6 (phoB) 0.68 130.6pTZ18R (vector) 0.60 2.9

a Strains of BW7656 (phoB23) containing the plasmids indicated were grownin AB minimal medium for 20 h at 37°C and assayed for AP activity (Millerunits). Results are averages of experiments done in quadruplicate.

RESULTS

Isolation of an A. tumefaciens clone that complements E. coliphoB. An A. tumefaciens cosmid library (18) was mobilized bytriparental mating (17) into JM105 phoB::Tn5 (43). Transcon-jugants were screened for AP activity on low-phosphate-concentration media (0.05 mM phosphate) containing thechromogenic substrate X-Phos. Two cosmids, pNM204 andpNM205, that restored phosphate starvation-inducible alkalinephosphatase activity were identified (data not shown). Thesecosmids did not confer AP activity on a strain deleted forphoA,the structural gene for AP. pNM204 and pNM205 wereanalyzed by endonuclease restriction digestion and were foundto share a 4.5-kb HindlIl restriction fragment. This fragmentwas cloned from pNM204 into the vector pTZ18R, creatingplasmid pNM139 (Fig. 1). pNM139 complemented JM105phoB::TnS (Fig. 1). This plasmid was used as a probe in aSouthern hybridization and was shown to hybridize to A.tumefaciens chromosomal DNA (data not shown).To localize the gene within pNM139 that complemented

phoB, deletions within the 4.5-kb Hindlll insert were con-structed by digestion with endonucleases that cut within themultiple cloning site of pTZ18R and within the insert DNA(Fig. 1). These constructs were then introduced by transfor-mation into JM105 phoB::Tn5 and screened for AP activity.The resulting plasmids displayed full, partial, or no comple-mentation ability (Fig. 1).Complementation of a phoB point mutation. The TnS inser-

tion within phoB is polar on phoR and thus complicatedanalysis of the complementation data. The plasmid pNM142,which contained one of the smallest fragments sufficient tocomplement the phoB::TnS mutation, was tested for the abilityto complement a phoB point mutation in a wild-type phoRbackground. The plasmid pNM142 (chvI) (Fig. 1), pBC6APstI(which contains the E. coli phoBR operon in pUC19) (57), orpTZ18R (the vector control) was introduced by transformationinto BW7656, which contains the phoB23 point mutation (57).These strains were grown in AB medium with an initialconcentration of 10 or 0.1 mM phosphate and assayed for APactivity. pNM144 restored phosphate-limiting AP expression,and the level of activity was 30 to 35% of the level in the straincontaining pBC6APstI (Table 2).

Nucleotide sequence of the A. tumefaciens phoB-like gene. A950-bp region of DNA required to complement phoB wassequenced. A single open reading frame was identified, and itencoded a predicted protein of 241 amino acids (Fig. 2). Thededuced sequence of this protein was compared with theprotein GenBank data base by using the FASTA program ofPearson and Lipman (45). The predicted protein had a strongsimilarity to the class of bacterial transcriptional regulatorswith 35% amino acid identity to the E. coli PhoB protein (36),36% identity to the B. subtilis PhoP protein (48), 35% identityto the Pseudomonas aeruginosa PhoB protein (2), and 35%identity to the Rhizobium meliloti PhoB (39). The putative

ACAACCTGGCTGGGTGAATGTCGGAGTTACGAGAAAATGACGATG

GAGACCAACTACATGCAGACGATCGCGCTTGTCGACGATGACCGGM Q T I A L V D D D R

AATATTCTCACTTCGGTATCGATCGCGCTCGAACGCGAAGGCTACN I L T S V S I A L E R E G Y

CGGGTCGAAACGTATACGGACGGCGCCTCCGCCGTTGACGGGTTGR V E T Y T D G A S A V D G L

ATCGCCCGTCCGCCGCAGCTTGCCATTTTCGATATTAAGATGCCCI A R P P Q L A I F D I K M P

CGCATGGACGGCATGGAGCTTTTGCAGCGCCTTCGCCAGAAATCGR M D G M E L L Q R L R Q K S

GATATTCCCGTTATCTTCCTCACCTCGAAGGATGAGGAGATCGACD I P V I F L T S K D E E I D

GAATTGTTCGGCCTGAAGATGGGTGCAGACGATTTCATCACAAAGE L F G L K M G A D D F I T K

CCGTTTTCGCAGCGCCTTCTGGTGGAGCGCGTCAAGGCCATCCTGP F S Q R L L V E R V K A I L

CGCCGTGCCAGCAGCCGTGAGGCGTCCGCTGCAACCGGCGGCACGR R A S S R E A S A A T G G T

CTGAAGCCGACCGCGGACCAGCAGGCGCGCACGCTGGAGCGTGGAL K P T A D Q Q A R T L E R G

CAGCTCGCCATGGACCAGGAACGCCACACCTGCACCTGGAAGGGTQ L A M D Q E R H T C T W K G

GAGCCGGTGACGCTGACAGTGACTGAATTCCTCATCCTGCATTCGE P V T L T V T E F L I L H S

CTTGCACCAGCGGCCGGGCGTGGTGAAAAGCGCGACGCGTTGATGL A P A A G R G E K R D A L M

GATGCCGCCTATGACGAACAGGTCTATGTGGACGACCGCACCATCD A A Y D E Q V Y V D D R T I

GACAGCCACATCAAGCGTCTTCGCAAGAAGTTCAAACTGGTCGATD S H I K R L R K K F K L V D

GGCGACTTCGATATGATTGAAACGCTTTATGGCGTAGGATATCGCG D F D M I E T L Y G V G Y R

TTCGCCGAAGCGGCTTAAAGAAGCCGACTCAAGCAGAAGGGACCGF A E A A *

CCCGAACCGGCTAAGGACTTTAATACGTGTTGAAGAAACGCCGGA

ACCGTCTCGACAGTGACGATGCCGAAGAACGCGCCGAACGCCGCC

ATCGTATCATCCAGCTGACGATCTACGCGCATTTCGCATCGGGTTT

45

90

135

180

225

270

315

360

405

450

495

540

585

630

675

720

765

810

855

900

946

FIG. 2. Nucleotide and predicted amino acid sequences of chvI.Underlining indicates a possible Shine-Dalgarno sequence, and theasterisk indicates a stop codon.

receiver and DNA binding domains of ChvI (44) were sepa-rated by 10 to 12 more amino acids than most homologs,including PhoB of E. coli (Fig. 3). This gene, for reasonsdiscussed below, was designated chvI and not phoB.

Directed mutagenesis of chvI and marker exchange into theA. tumefaciens chromosome. A cassette encoding resistance tokanamycin was inserted into the unique EcoRI site within chvI,as described in Materials and Methods. This disrupted allelewas cloned into the suicide vector pKNG101 (31) to createpNM156. The vector pKNG101 encodes resistance to strepto-

VOL. 175, 1993

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

6630 MANTIS AND WINANS

ChvI

PhoB

ChvI

PhoB

10 20 30 40 50MQTIALVDDDRNILTSVSIALEREGYRVETYTDGASAVDGLIARPPQLAIFDIKM

MARRILVVEDEAPIREMVCFVLEQNGFQPVEAEDYDSAVNQLNEPWPDLILLDWML10 20 30 40 50

60 70 80 90 100 110PRMDGMELLRRLRQKS---DIPVIFLTSKDEEIDELFGLKMGADDFITKPFSQRLLVERV

PGGSGIQFIKHLKRESMTRDIPVVMLTARGEEEDRVRGLETGADDYITKPFSPKELVARI60 70 80 90 100 110

120 130 140 150 160 170ChvI KAILRRASSR--EASAATGGTLKPTADQQARTLERGQLAMDQERHTCTWKGEPVTLTVTE

::..:: :.. :. ..:. :...: :... :PhoB KAVMRRISPMAVEE-----------------VIEMQGLSLDPTSHRVMAGEEPLEMGPTE

120 130 140 150

ChvI

PhoB

ChvI

PhoB

180 190 200 210 220 230FLILHSLAPAAGRGEKRDALMDAAYDEQVYVDDRTIDSHIKRLRKKFKLVDGDFDMIETL

FKLLHFFMTHPERVYSREQLLNHVWGTNVYVEDRTVDVHIRRLRKALE-PGGHDRMVQTV170 180 190 200 210

240YGVGYRFAEAA

RGTGYRFSTRF220

FIG. 3. Optimal alignment of ChvI of A. tumefaciens with PhoB of E. coli, as obtained from the FASTA program. Colons indicate identicalresidues, and periods indicate conservative substitutions. A 12-amino-acid gap was added to PhoB to obtain optimal alignment.

mycin, has a conditional origin of replication, and has the sacBgene which mediates sucrose sensitivity to A. tumefaciens (31).Plasmid pNM156 was mobilized into A. tumefaciens strains byconjugation (50), and transconjugants were plated on LB-kanamycin-streptomycin plates. Colonies that arose had plas-mid pNM156 integrated in the chromosome at the chvI locusby a single homologous recombination event (Fig. 4). To selectfor strains which had undergone a double-recombination eventin which the mutated allele replaced the wild-type allele, theseKmr Smr colonies were pooled and plated on AB-sucrose(5%)-kanamycin plates. Approximately 40% of the Kmr Sucrcolonies were sensitive to streptomycin, suggesting that theyhad lost the plasmid DNA and undergone marker replace-ment. Chromosomal DNA from 11 of these strains was iso-lated and analyzed by Southern blot hybridization withpNM144 as a probe (Fig. 4). All 11 strains had a Sall restrictionpattern that agreed with the pattern expected if the disruptedchvI allele had replaced the wild-type allele (Fig. 4). Thismutation was constructed in isogenic strains of A. tumefacienscontaining the Ti plasmid or lacking the Ti plasmid, thuscreating NM104 or NM103, respectively.PhoB or not PhoB? Phosphate regulatory gene mutations in

E. coli, P. aeruginosa, and B. subtilis abolish phosphate starva-tion-inducible phosphatase activity. A. tumefaciens has a phos-phate starvation-inducible phosphatase activity (Table 3). Todetermine the effect of the chvI mutation on A. tumefaciensphosphatase activity, NM104 was cultured in AB media withhigh (10 mM) or low (0.1 mM) phosphate concentration andassayed for phosphatase activity (Table 3). NM104 showedphosphate-limiting phosphatase expression and had activitylevels similar to those of the wild type. Furthermore, NM104grew as well as the wild type in phosphate-limiting media (10mM to 0.01 mM) (data not shown). Thus, the chvI mutationdid not produce the phenotype expected for a mutation in a

phosphate regulatory gene. To test the effect of this mutationon phosphate starvation-inducible virG expression, the plasmidpSW174, which contains a virG::lacZ fusion, was transformedinto NM104. NM104(pSW174) and A348(pSW174) weregrown in media with high or low phosphate concentration andassayed for 3-galactosidase activity. The mutant and the wildtype had identical levels of ,B-galactosidase activity when grownin phosphate-limiting media (Table 3). We conclude that chvIis not required to induce phosphatase activity or virG inresponse to phosphate starvation.

Characterization of NM104. We were unable to isolate chvIinsertion mutants on LB medium, but we could isolate suchmutants on AB minimal medium (see above). This suggestedthat LB medium may be toxic to the chvI mutant. The growthcurves of A348 and NM104 in AB medium (a minimal saltsmedium) or in LB medium (tryptone [1% wt/vol], yeast extract[0.5% wt/vol], and NaCl [1% wt/vol]) were compared. NM104and A348 had identical growth curves in AB medium (Fig. 5).However, NM104 stopped growing in LB broth within 1generation. The ability of NM104 to grow in LB was restoredwhen chvI was supplied in trans on the plasmid pNM157 (Fig.5). NM104 did not grow in LB containing excess phosphate(100 mM) or in AB containing LB (10% vol/vol). The individ-ual components of LB were added to AB, and these mediawere inoculated with NM104. The mutant did not grow in ABplus tryptone (1% wt/vol) or in AB plus peptone (1% wt/vol),it grew poorly in AB plus Casamino Acids (1% wt/vol), and itgrew well in AB plus yeast extract (0.5% wt/vol). To determinewhether amino acids inhibited growth of NM104, AB minimalmedium was supplemented with each individual amino acid(15) or all 20 individual amino acids combined and inoculatedwith NM104. NM104 grew normally in these media. Differentcombinations of amino acid concentrations were not tested.NM104 was hypersensitive to acidic culture media. The

J. BACTERIOL.

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

chvI OF A. TUMEFACIENS 6631

d

C

ba

b a8IonsI I hi

lane 1 = ~

a a

4, dlane 2

lane 3+4

4'c

chromosome

lasmid

C8 8 Km'I I

kf aS SI I

aS s

-. I IFIG. 4. Southern blot analysis of chvI marker exchange replacement events. (A) Chromosomal DNA was digested with Sall, size fractionated

by gel electrophoresis, and hybridized to biotinylated pNM142, as described in Materials and Methods. Bands appearing on the autoradiographare labelled at left with lowercase letters that correspond to the letters labelling the fragments indicated on the schematic diagram in panel B.Lanes: 1, wild-type A348; 2, A348 in which pNM156 is integrated by a single homologous recombination event; 3 and 4, replacement of thekanamycin gene within chvl. (B) Schematic depiction of the chromosome arrangement at the chvI locus shown in panel A. S, SailI restriction sites.

growth curves of NM104 and A348 grown in AB media at pH7.0, 6.0, 5.5, or 5.0 were determined. A348 grew well in allmedia, although the rate of cell doubling decreased as the pHof the medium decreased (Fig. 6B). NM104 grew poorly at pH

TABLE 3. Phosphate starvation-inducible phosphatase activity andvirG::iacZ activity in NM104'

Activity at phosphate concn:Activity and strain

10 mM 0.1 mM

PhosphataseNM104 0.39 23A348 0.20 34

,B-GalactosidaseNM104(pSW174) 4.6 177A348(pSW174) 6.0 164a Cultures were grown in AB minimal medium containing 10 or 0.1 mM

phosphate, as described in Materials and Methods, and assayed for AP activityafter 19 h or measured for 13-galactosidase activity after 17 h (Miller units).Plasmid pSW174 contains a virG::IacZ fusion (59). Results are averages ofexperiments done in quadruplicate.

5.5 and was unable to grow at pH 5.0 (Fig. 6A). We alsoobserved that NM104 was hypersensitive to the antibioticstetracycline and carbenicillin (data not shown).Tumorigenesis assays and vir gene induction in NM104. To

determine whether chvI was required for tumorigenesis, Kdiagremontiana leaves were inoculated with NM104, as de-scribed in Materials and Methods. NM104 did not causetumors on K diagremontiana (Fig. 7B). Avirulence may be dueto the inability to induce the vir genes, the inability to survivethe plant wound environment, or a combination of both. Totest vir gene induction, NM104 was transformed with pSW174,a broad-host-range plasmid containing a virG::lacZ fusion, orpSW219, a broad-host-range plasmid with a virB::lacZ fusion.These strains were grown in AB medium, supplemented withAS and buffered at pH 7.0, 6.0, or 5.5, and assayed for,-galactosidase activity. There was no increase in 3-galactosi-dase activity in NM104(pSW174) grown under these condi-tions (Fig. 8A), whereas 3-galactosidase activity increasedmore than 100-fold in A348(pSW174) grown at pH 5.5 with100 ,uM AS. These data suggested that virG expression wasabolished in the chvI background. virB expression was alsosignificantly attenuated: 3-galactosidase activity in NM104

nomossommooLm I 1-

VOL. 175, 1993

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

6632 MANTIS AND WINANS

1

00

0._0

c

C._0.

QCL

0.1

0.010

iA

4 8 12

Time (Hours)FIG. 5. Growth of NM104 in LB and AB media. A. tumefaciens

A348 (wild type), NM104 (chvI), and NMI04(pNM157) (chvI+) weregrown overnight in AB minimal medium to mid-log phase and thenshifted to LB (shaded symbols) or AB (open symbols) medium.Plasmid pNM157 is a broad-host-range plasmid which contains the4.5-kb HindIll fragment that encodes chvI. A, A348; *, NM104; 0,NM104(pNM157).

(pSW219) grown at pH 5.5 with 100 F.M AS, optimal inducingconditions, was less than 10% of the wild-type activity (Fig.8B).As described in the introduction, virG is transcribed from

two promoters, P1 and P2. The P2 promoter is induced byacidic culture media, independent of VirA-VirG (37, 59).

0

0

0

4c

In

0

0.1

0.01

FIG. 7. Tumorigenesis assay of chvl mutant on K diagremontiana.A. tumefaciens strains were grown overnight in AB medium, and a 1-mlvolume of each culture (ca. 10 CFU/ml) was used to inoculatewounded K diagremontiana leaves as described in Materials andMethods. Panels are as follows: A, A348 (wild type); B, NM104 (chvImutant); and C, A136 (lacking the Ti plasmid). Leaves were photo-graphed 5 weeks postinoculation.

1-Galactosidase activity was measured in strain NM 104(pSW174) or A348(pSW174) grown in AB media buffered atpH 7.0, 6.0, or 5.5. A 7- to 10-fold induction of 3-galactosidaseactivity was observed in A348(pSW174) grown at pH 5.5compared with the activity at pH 7.0, which agrees withprevious results (37). No increase in 3-galactosidase activity

1

0.1

0.010 3 6 9 12 0 3 6 9 12

Time (h)FIG. 6. Growth of NM 104 in acidic culture media. A. tumefaciens NM104 (chvI) (A) or A348 (wild type) (B) were grown in AB medium (pH

7.0) to mid-log phase and then diluted into AB medium buffered at pH 7.0 (A), pH 6.0 (0), pH 5.5 (*), or pH 5.0 (H).

J. BACTERIOL.

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

chvI OF A. TUMEFACIENS 6633

bO

45

30

15

07 6 5.5

7.0 6.0 5.5

1800

1200

600

0p

pH

7 6 5.5

7.0 6.0

FIG. 8. vir gene induction by AS and acidic pH in a chvI mutant. A. tumefaciens strains were cultured in AB media buffered to the pHs indicatedand assayed for 3-galactosidase activity between 16 and 20 h. AS (100 FtM) was added to the cultures whose results are shown in panels A andB. Panels: A and C, NM104(pSW174) (solid bars) and A348(pSWI74) (hatched bars); B, NMI04(pSW219) (solid bars) and A348(pSW219)(hatched bars); D, NM103(pCC101) (solid bars) and A136(pCC101) (hatched bars). pSW174 contains virG::lacZ, pSW219 contains virB::lacZ, andpCC101 contains Plac::lacZ. The numbers are averages of experiments done in quadruplicate. ND, not determined. The ,B-galactosidase activityobserved from A348(pSW219) (virB::1acZ) agrees with previous results (13).

was observed in NMI04(pSW174) grown in acidic culturemedia (Fig. 8C).We were concerned that the lack of vir gene expression in a

chvI mutant background was due simply to the poor growth ofthis strain in acidic culture media (Fig. 6). Plasmid pCC101contains the E. coli lacZ gene fused to the lac promoter, whichis expressed strongly and constitutively in A. tumefaciens (13).This plasmid was introduced by transformation into NM103,and 3-galactosidase activity was measured from cells grown inmedia buffered at pH 7.0 or pH 5.5. A136(pCCIO1) hadidentical levels of 3-galactosidase activity when grown atpH 7.0 and at pH 5.5. 3-Galactosidase activity fromNM103(pCC101) grown at pH 5.5 was 80% of the activity fromcells grown at pH 7.0.The virG P2 promoter is induced by acidic culture media and

is thought to function in increasing intracellular concentrationsof VirG to allow efficient vir gene induction (59). Since P2expression was abolished in NM104 (Fig. 8C), we hypothesizedthat this may result in a low intracellular level of VirG, whichcould in turn attenuate vir gene induction. To test this, we

constructed a strain in which virG was constitutively expressedfrom the lac promoter. NM103 was transformed with theplasmid pCH116, which contains a Plac::virG fusion and a

virB::lacZ fusion (10), and with plasmid pCH114, which con-tains virA (10). NM103(pCH114)(pCH116) and A136(pCH114)(pCH116) were grown in induction media bufferedat pH 7.0, 6.0, or 5.5 and assayed for 3-galactosidase activity.There was no detectable increase in 3-galactosidase activity inNM103(pCH114)(pCH116) under these conditions (data notshown). Therefore, expression of virG from the lac promoterdid not rescue vir gene expression.

Sensitivity of chvI mutants to plant wound sap. Avirulenceof chvI mutants may be due, in part, to the inability to survive

in the wound environment, which is usually acidic and containsa number of antimicrobial activities, including low-molecular-weight compounds called phytoalexins (32). We developed a

simple assay to determine whether the plant secreted com-

pounds that could inhibit growth of NM104. K diagremontianaleaf sections were placed on AB plates that were overlaid witha 0.7% top agar containing A. tumefaciens NM104 (chvI) orA348 (wild type). A zone of growth inhibition (ca. 30 mm indiameter) was observed around the plant slice on platesoverlaid with NM104 (Fig. 9), while no such zone was found on

plates overlaid with A348 (Fig. 9). The pH of the mediumsurrounding the plant slice did not change, as determined bythe addition of the pH indicator dye bromocresol blue (datanot shown). This was expected, since AB medium is bufferedwith 25 mM phosphate. Using the same assay method, we

tested other A. tumefaciens chromosomal mutations that affectvirulence, including chvA, chvB, and chvE, and two exopolysac-charide mutants, exoA and exoC. These mutants all had thesame phenotype as the wild type, except for exoC, whichactually showed enhanced growth around the plant slice (datanot shown). We hypothesized that the chvI mutant may besensitive to phytoalexins. Three purified phytoalexins, glyceol-lin and gossypol from Phaseolus vulgaris and 6-methoxymellienfrom carrot roots, were spotted on plates overlaid with top agarcontaining A348 or NM104. These phytoalexins did not inhibitgrowth of A348 or NM104 (data not shown).

DISCUSSION

A. tumefaciens must be metabolically active to efficientlyinfect plant hosts (21, 34); growth at the plant wound site mayrequire the coordinate regulation of genes involved in metab-olism and/or genes necessary to overcome plant defense

620

465

310

I1U

-b

o =

l

ma

155

0

45

30

15

05.5

VOL. 175, 1993

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

6634 MANTIS AND WINANS

FIG. 9. Sensitivity of the chvI mutant to K diagremontiana woundsap. AB plates were overlaid with a 0.7% AB top agar seeded with A.tumefaciens A348 (top) or NM104 (bottom). The youngest full-sizeleaves from a K diagremontiana plant were removed, cut with a razorblade to obtain cross sections (2 to 5 mm thick), and then placed on thesolidified top agar. Plates were incubated at 30°C and photographedafter 48 h.

mechanisms. In this paper, we have identified an A. tumefa-ciens chromosomal response regulatory gene, chvI, necessaryfor tumorigenesis on K diagremontiana (Fig. 7). chvI wasisolated from an A. tumefaciens cosmid library as a clone thatcomplemented an E. coli phoB mutation (Fig. 1 and Table 2),although chvI was not required for phosphatase regulation inA. tumefaciens (Table 3). chvI was, however, required forgrowth in complex (Fig. 5) or acidic media (Fig. 6), suggestingthat chvI may play a central role in the regulation of genesinvolved in metabolism or adaptation to certain environmentalconditions. Furthermore, in a chvI null background, vir gene(virG and virB) induction in response to AS and virG inductionin response to acidic pH were significantly attenuated (Fig. 8Athrough C), but expression of a Plac::lacZ fusion under thesame conditions was not (Fig. 8D). In a separate study, a genecalled chvG was identified immediately downstream of chvI,

and it encodes a protein with amino acid similarity to thefamily of bacterial histidine kinases (12). Thus, it is likely thatChvG and ChvI constitute a two-component regulatory system,the first chromosomal two-component system identified in A.tumefaciens, which is directly (e.g., in a regulatory hierarchy) orindirectly (e.g., by metabolic regulation) required for vir geneexpression and virulence.To complement aphoB mutation, ChvI must receive a signal

in response to phosphate starvation, recognize phosphate-regulated promoters, and activate transcription of these pro-moters (56). ChvI is probably activated by phosphorylation, asare the other members of this protein family, such as VirG andPhoB (29, 35). ChvI contains an aspartyl residue that isconserved within response regulators, which in some cases hasbeen shown to be the site of phosphorylation (44). In E. coli,PhoR mediates PhoB phosphorylation and dephosphorylation,and we expect that it mediates ChvI in a similar way. ChvIprobably activated AP activity (Table 2) by activating transcrip-tion of phoA. However, we have no evidence that ChvIrecognizes the same DNA sequence as PhoB or that ChvIactivated other PhoB-dependent promoters. Despite the sim-ilarity of the virG PI promoter to the consensus pho boxsequence (59) and evidence that the P1 promoter of virG fromAgrobacterium rhizogenes is transcribed in E. coli in a PhoB-dependent manner (3), a virG::lacZ fusion was not phosphateregulated in an E. coli host containing chvI (38).

Evidence suggests that chvI is not the primary phosphateregulatory gene in A. tumefaciens. First, the amino acid identity(35%) between ChvI and PhoB of E. coli (Fig. 3) was muchlower than the identity between PhoB of E. coli and otherPhoB-like proteins, including PhoB of P. aeruginosa (59%) (2),PhoB of Rhizobium meliloti (47%) (39), or PhoP of B. subtilis(40%) (48). Second, and more importantly, the phenotype ofthe chvI mutant did not resemble the phenotype expected fora strain with a mutation in a phosphate regulatory gene (Table3). In other bacteria in which PhoB homologs have beenidentified, phoB mutations abolish phosphate starvation-induc-ible phosphatase activity (1, 2, 39, 48, 56). We do not excludethe possibilities that chvI may be involved in some aspect ofphosphate regulation in A. tumefaciens and that we have notdetected its role by measuring only phosphatase activity.Moreover, A. tumefaciens phosphate regulation may be morecomplex than initially expected. It was previously suggestedthat A. tumefaciens may have multiple phosphatases which maybe differentially regulated (9).The chvI insertion mutation was pleiotropic. We note that

this pleiotropy may be due to a polar effect on genes down-stream of chvI (Fig. 4). However, any gene affected by thispolarity must be located within the 4.5-kb HindlIl fragmentthat includes chvI (Fig. 1), because this fragment was sufficientto complement the chvI mutation, at least for growth incomplex media (Fig. 5). chvG mutants have a phenotypeidentical to that of the chvI mutant (12). The inability ofNM104 to grow in complex media explains why we were unableto obtain chvI mutations on LB (see Results). Our dataindicated that peptone and tryptone media were toxic toNM104 rather than being limiting for an essential nutrient(e.g., phosphate), because addition of these compounds tominimal media arrested growth (Fig. 5). Addition of yeastextract to minimal media did not affect growth of NM104 (datanot shown), although on the basis of analyses done by DifcoLaboratories, yeast extract has a chemical composition similarto that of peptone and tryptone. It is interesting that the ArcAresponse regulator in E. coli is a pleiotropic regulatory proteinrequired for expression of F plasmid tra genes, some of whichare homologous to vir genes (28, 58).

J. BACTERIOL.

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

chvl OF A. TUMEFACIENS 6635

Acidic pH-sensitive mutants have been identified in two

gram-negative species, R. meliloti and Salmonella typhimurium(20, 24). In R. meliloti, four classes of Tn5 mutants that are

unable to grow or maintain a neutral intracellular pH at an

external pH of 5.6 were isolated (24). These mutants were not

tested for the ability to form nodules on plant hosts. In S.typhimurium, two loci, unc and phoP, which, when mutated,cause an acid-sensitive phenotype have been identified. Theunc operon encodes the ATPase (F0F1) involved in H+translocation and possibly pH homeostasis (20). phoP encodesa response regulatory protein homologous to PhoB and isrequired for virulence and resistance to host macrophage-produced cationic peptides, called defensins (19, 26, 42). Inaddition to possessing the acid-sensitive phenotype, the chvImutant was also found to be hypersensitive to carbenicillin anddetergents like sodium dodecyl sulfate (12a). One possibleexplanation for this phenotype is that the mutant may be more

permeable to certain compounds than the wild type. In neu-

trophilic bacteria, like A. tumefaciens, the regulation of cyto-plasmic pH, and thus resistance to high concentrations ofextracellular protons, is dependent on the intrinsic imperme-ability of the lipid bilayer (6).

Acidic culture medium influences A. tumefaciens vir gene

expression through at least two independent regulatory path-ways, and both of these pathways were impaired in a chvl nullbackground (Fig. 8). The first pathway involves vir gene

induction, which occurs optimally when A. tumefaciens iscultured in media between pH 5.5 and pH 5.0 and does notoccur at neutral pH. The second pathway involves the P2promoter of virG, which is transcriptionally induced by acidicpH by a chromosomal regulatory system, independent ofVirA-VirG (37, 59). Attenuated vir gene expression in the chvImutant was not due entirely to the poor growth of that strainin acidic culture media because (i) NM103 containing a

plasmid with a plac::lacZ fusion grown at pH 5.5 had 80% ofthe 3-galactosidase activity of the wild type grown under thesame conditions (Fig. 8D) and (ii) in suboptimal inducingconditions (e.g., pH 6.0) under which NM104 grew reasonablywell (Fig. 6), there was still significant reduction of vir expres-

sion (Fig. 8A and B). The pH requirement for vir gene

induction is thought to occur in a step of the signal transduc-tion pathway prior to VirG phosphorylation, since VirG mu-

tants that bypass this pH requirement were isolated (27). TheP2 promoter is thought to be necessary to maintain theintracellular pool of VirG required for efficient vir gene

induction. In the chvl mutant, constitutive expression of virGfrom the lac promoter was not sufficient to restore virBinduction, indicating that the lack of vir gene expression was

not due to limiting VirG concentrations (data not shown).Using a simple bioassay, we found that the chvI mutant was

unable to grow in the vicinity of a K diagremontiana leaf slice(Fig. 9), suggesting that something secreted from the plantslice was toxic for growth of the chvl mutant. Plants have a

number of wound-inducible and pathogen-inducible defenseresponses (32), which include the production of low-molecu-lar-weight diffusible compounds with antimicrobial activitycalled phytoalexins. We tested several purified phytoalexins forthe ability to inhibit the growth of a chvI mutant; none hadantibacterial activity (38). We have not tested phytoalexinsfrom K diagremontiana, nor have we tested other plants, forthe ability to inhibit growth of NM104. It is tempting tospeculate that K diagremontiana may secrete one or more

metabolites toxic to A. tumefaciens and that chvI is required forresistance to these compounds. An elucidation of the environ-mental signals that activate ChvG-ChvI, and the identification

of ChvI target genes, may shed light on the role these proteinsplay in the regulation of A. tumefaciens virulence.

ACKNOWLEDGMENTS

We thank Trevor Charles and Eugene Nester for sharing resultsprior to publication. We thank Anna-Maria Torriani for E. coli JM105phoB::Tn5, Barry Wanner for E. coli phoB strains and the cloned phoBgene, Steve Farrand for the cosmid library, Valley Stewart for plasmidpVJS303, Steve Zinder for assistance with the Southern blot protocol,and Greg DiCenzo for purified phytoalexins. Thanks to Clay Fuquaand Barry Wanner for helpful discussions and technical advice.

This work was supported by NIH grant no. I R29 GM2893-01.

REFERENCES1. Aiba, H., M. Nagaya, and T. Mizuno. 1993. Sensor and regulator

proteins from the cyanobacterium Synechococcus speciesPCC7942 that belong to the bacterial signal-transduction proteinfamilies: implications in the adaptive response to phosphatelimitation. Mol. Microbiol. 8:81-91.

2. Anba, J., M. Bibaud, M. L. Vasil, and A. Lazdunski. 1990.Nucleotide sequence of the Pseudomonas aeruginosa phoB gene,the regulatory gene for the phosphate regulon. J. Bacteriol.172:4685-4689.

3. Aoyama, T., M. Takanami, K. Makino, and A. Oka. 1991. Cross-talk between the virulence and phosphate regulons of Agrobacte-rium tumefaciens caused by an unusual interaction of the transcrip-tional activator with a regulatory DNA element. Mol. Gen. Genet.227:385-390.

4. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. A. Smith,J. G. Seidman, and K. Struhl. 1987. Current protocols in molec-ular biology. John Wiley & Sons, Inc., New York.

5. Binns, A. N., and M. F. Thomashow. 1988. Cell biology ofAgrobacterium infection and transformation of plants. Annu. Rev.Microbiol. 42:575-606.

6. Booth, I. R. 1985. Regulation of cytoplasmic pH in bacteria.Microbiol. Rev. 49:359-378.

7. Brickman, E., and J. Beckwith. 1975. Analysis of the regulation ofEscherichia coli alkaline phosphatase synthesis using deletions and480 transducing phages. J. Mol. Biol. 96:307-316.

8. Cangelosi, G. A., R. G. Ankenbauer, and E. W. Nester. 1990.Sugars induce the Agrobacterium virulence genes through aperiplasmic binding protein and a transmembrane signal protein.Proc. Natl. Acad. Sci. USA 87:6708-6712.

9. Cangelosi, G. A., E. A. Best, G. Martinelli, and E. W. Nester. 1991.Genetic analysis of Agrobacterium. Methods Enzymol. 204:384-397.

10. Chang, C.-H., and S. C. Winans. 1992. Functional roles assigned tothe periplasmic, linker, and receiver domains of the Agrobacteriumtumefaciens VirA protein. J. Bacteriol. 174:7033-7039.

11. Charles, T. C., S. Jin, and E. W. Nester. 1992. Two-componentsensory transduction systems in phytobacteria. Annu. Rev. Phyto-pathol. 30:463-484.

12. Charles, T. C., and E. W. Nester. 1993. A chromosomally encodedtwo-component sensory transduction system is required for viru-lence of Agrobacterium tumefaciens. J. Bacteriol. 175:6614-6625.

12a.Charles, T. C., and E. W. Nester. Personal communication.13. Chen, C.-Y., and S. C. Winans. 1991. Controlled expression of the

transcriptional activator gene virG in Agrobacterium tumefaciensby using the Escherichia coli lac promoter. J. Bacteriol. 173:1139-1144.

14. Cooley, M. B., M. R. D'Sousa, and C. I. Kado. 1991. virC and virDoperons of the Agrobacterium Ti plasmid are regulated by the roschromosomal gene: analysis of the cloned ros gene. J. Bacteriol.173:2608-2616.

15. Davis, R., D. Botstein, and J. Roth. 1989. Advanced bacterialgenetics. Cold Spring Harbor Laboratory, Cold Spring Harbor,N.Y.

16. Dente, L., G. Cesareni, and R. Cortese. 1983. pEMBL: a newfamily of single stranded plasmids. Nucleic Acids Res. 11:1645-1655.

17. Ditta, G., S. Stanfield, D. Corbin, and D. Helinski. 1980. Broadhost range DNA cloning system for gram-negative bacteria: con-

VOL. 175, 1993

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

6636 MANTIS AND WINANS

struction of a gene bank of Rhizobium meliloti. Proc. Natl. Acad.Sci. USA 77:7374-7351.

18. Farrand, S. K., S. P. O'Morchos, and J. McCutchan. 1989.Construction of an Agrobacterium tumefaciens C58 recA mutant. J.Bacteriol. 171:5314-5321.

19. Fields, P. I., E. A. Groisman, and F. Heffron. 1989. A Salmonellalocus that controls resistance to microbial proteins from phago-cytic cells. Science 243:1059-1062.

20. Foster, J. W., and H. K. Hall. 1990. Adaptive acidification toler-ance response of Salmonella typhimurium. J. Bacteriol. 172:771-778.

21. Garfinkel, D. J., and E. W. Nester. 1980. Agrobacterium tumefa-ciens mutants affected in crown gall tumorigenesis and octopinecatabolism. J. Bacteriol. 144:732-743.

22. Gay, P., D. Le Coq, M. Steinmetz, T. Berkelman, and C. I. Kado.1985. Positive selection for entrapment of insertion sequenceelements in gram-negative bacteria. J. Bacteriol. 164:918-921.

23. Gelvin, S. B. 1990. Crown gall disease and hairy root disease: asledgehammer and a tackhammer. Plant Physiol. 92:281-285.

24. Goss, T. J., G. W. O'Hara, M. J. Dilworth, and A. R. Glenn. 1990.Cloning, characterization, and complementation lesions causingacid sensitivity in TnS-induced mutants of Rhizobium melilotiWSM419. J. Bacteriol. 172:5173-5179.

25. Gray, J., J. Wang, and S. B. Gelvin. 1992. Mutation in the miaAgene of Agrobacterium tumefaciens results in reduced vir geneexpression. J. Bacteriol. 174:1086-1098.

26. Groisman, E. A., E. Chiao, C. J. Lipps, and F. Heffron. 1989.Salmonella typhimurium phoP virulence gene is a transcriptionalregulator. Proc. Natl. Acad. Sci. USA 86:7077-7081.

27. Han, D. C., C.-Y. Chen, Y.-F. Chen, and S. C. Winans. 1992.Altered-function mutations in the transcriptional regulatory genevirG of Agrobacterium tumefaciens. J. Bacteriol. 174:7040-7043.

28. luchi, 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.

29. Jin, S., R. K. Prusti, T. Roitsch, R. G. Akenbauer, and E. W.Nester. 1990. Phosphorylation of the VirG protein of Agrobacte-rium tumefaciens by the autophosphorylated VirA protein: essen-tial role in biological activity of VirG. J. Bacteriol. 172:4945-4950.

30. Kado, C. I. 1991. Molecular mechanisms of crown gall tumorigen-esis. Crit. Rev. Plant Sci. 10:1-32.

31. Kaniga, K., I. Delor, and G. R. Cornelis. 1991. A wide-host-rangesuicide vector for improving reverse genetics in Gram-negativebacteria: inactivation of the blaA gene of Yersinia enterocolitica.Gene 109:137-141.

32. Lamb, C. J., M. A. Lawton, M. Dron, and R. A. Dixon. 1989.Signals and transduction mechanisms of plant defenses againstmicrobial attack. Cell 56:215-224.

33. Leroux, B., M. F. Yanofsky, S. C. Winans, J. E. Ward, S. F. Zeigler,and E. W. Nester. 1987. Characterization of the virA locus ofAgrobacterium tumefaciens: a transcriptional regulator and hostrange determinant. EMBO J. 6:849-856.

34. Lippincott, J. A., and B. B. Lippincott. 1975. The genus Agrobac-terium and plant tumorigenesis. Annu. Rev. Microbiol. 44:377-405.

35. 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 phosphotransfer be-tween PhoR and PhoB proteins. J. Mol. Biol. 210:551-559.

36. Makino, K., H. Shinagawa, M. Amemura, and A. Nakata. 1986.Nucleotide sequence of the phoB gene, the positive regulatorygene for the phosphate regulon of Escherichia coli K-12. J. Mol.Biol. 190:37-44.

37. Mantis, N. J., and S. C. Winans. 1992. The Agrobacteriumtumefaciens vir gene transcriptional activator virG is transcription-ally induced by acidic pH and other stress stimuli. J. Bacteriol.174:1189-1196.

38. Mantis, N. J., and S. C. Winans. Unpublished data.39. McLean, P. A., C.-M. Liu, C. C. Sookdeo, and F. C. Cannon.

Genetics and phoB regulation of phosphonate utilization in Rhi-zobium meliloti 1021. Submitted for publication.

40. Metts, J., J. West, S. H. Doares, and A. G. Mattysse. 1991.Characterization of three Agrobacterium tumefaciens avirulent

mutants with chromosomal mutations that affect induction of virgenes. J. Bacteriol. 173:1080-1087.

41. Miller, J. H. 1972. Experiments in molecular genetics. Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.

42. Miller, S. I., A. M. Kukral, and J. J. Mekalanos. 1989. A twocomponent regulatory system (phoP phoQ) controls Salmonellatyphimurium virulence. Proc. Natl. Acad. Sci. USA 86:5054-5058.

43. Pahel, G., and A. Torriani. Unpublished data.44. Parkinson, J. S., and E. C. Kofoid. 1992. Communication modules

in bacterial signalling proteins. Annu. Rev. Genet. 26:71-112.45. Pearson, W. R., and D. J. Lipman. 1988. Improved tool for

biological sequence comparison. Proc. Natl. Acad. Sci. USA85:2444-2448.

46. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring Harbor Labo-ratory, Cold Spring Harbor, N.Y.

47. Sciaky, D., A. L. Montoya, and M.-D. Chilton. 1977. Fingerprintsof Agrobacterium Ti plasmids. Plasmid 1:238-253.

48. Seki, T., H. Yoshikawa, and H. Saito. 1987. Cloning and nucleotidesequence of phoP, the regulatory gene for alkaline phosphataseand phosphodiesterase in Bacillus subtilis. J. Bacteriol. 169:2913-2916.

49. Shimoda, N., A. Toyoda-Yamamoto, J. Nagamine, S. Usami, M.Katayama, Y. Sakagami, and Y. Machida. 1990. Control ofexpression of Agrobacterium vir genes by synergistic actions ofphenolic signal molecules and monosaccharides. Proc. Natl. Acad.Sci. USA 90:6684-6688.

50. Simons, R. W., U. Priefer, and A. Puhler. 1983. A broad host rangemobilization system for in vivo genetic engineering: transposonmutagenesis in gram negative bacteria. Bio/Technology 1:784-791.

51. Stachel, S. E., E. Messens, M. Van Montagu, and P. Zambryski.1985. Identification of the signal molecules produced by woundedplant cells that activate T-DNA transfer in Agrobacterium tumefa-ciens. Nature (London) 318:624-629.

52. Stachel, S. E., E. W. Nester, and P. C. Zambryski. 1986. A plantcell factor induces Agrobacterium tumefaciens vir gene expression.Proc. Natl. Acad. Sci. USA 83:379-383.

53. Stachel, S. E., and P. C. Zambryski. 1986. virA and virG control theplant-induced activation of the T-DNA transfer process of A.tumefaciens. Cell 46:325-333.

54. Veluthambi, K., R. K. Jayaswal, and S. B. Gelvin. 1987. Virulencegenes A, G, D, mediate the double stranded border cleavage ofT-DNA from Agrobacterium Ti-plasmid. Proc. Natl. Acad. Sci.USA 84:1881-1885.

55. Vernade, P., A. Herrera-Estralla, K. Wang, and M. Van Montagu.1988. Glycine betaine allows enhanced induction of the Agrobac-terium tumefaciens vir genes by acetosyringone at low pH. J.Bacteriol. 170:5822-5829.

56. Wanner, B. 1987. Phosphate regulation of gene expression inEscherichia coli, p. 1318-1325. In F. C. Neidhardt, J. L. Ingraham,K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger(ed.), Escherichia coli and Salmonella typhimurium: cellular andmolecular biology. American Society for Microbiology, Washing-ton, D.C.

57. Wanner, B. L., and B. D. Chang. 1987. The phoBR operon inEscherichia coli K-12. J. Bacteriol. 169:5569-5574.

58. Winans, S. C. 1992. Two-way chemical signaling in Agrobacterium-plant interactions. Microbiol. Rev. 56:12-31.

59. Winans, S. C. 1990. Transcriptional induction of anAgrobacteriumregulatory gene at tandem promoters by plant-released phenoliccompounds, phosphate starvation, and acidic growth media. J.Bacteriol. 172:2433-2438.

60. Winans, S. C., P. R. Ebert, S. E. Stachel, M. P. Gordon, and E. W.Nester. 1986. A gene essential for Agrobacterium virulence ishomologous to a family of positive regulatory loci. Proc. Natl.Acad. Sci. USA 83:8278-8282.

61. Winans, S. C., R. A. Kerstetter, and E. W. Nester. 1988. Transcrip-tional regulation of the virA and virG genes of Agrobacteriumtumefaciens. J. Bacteriol. 170:4047-4054.

62. Winans, S. C., R. A. Kerstetter, J. E. Ward, and E. W. Nester. 1989.A protein required for transcriptional regulation ofAgrobacteriumvirulence genes spans the cytoplasmic membrane. J. Bacteriol.171:1616-1622.

J. BAC-FERIOL.

on Novem

ber 23, 2020 by guesthttp://jb.asm

.org/D

ownloaded from