Vol. 176, No. 22JOURNAL OF BACrERIOLOGY, Nov. 1994, p. 7017-70230021-9193/94/$04.00+00Copyright X) 1994, American Society for Microbiology

Effects of rpoA and cysB Mutations on Acid Induction ofBiodegradative Arginine Decarboxylase in Escherichia coli

XIAOLU SHI AND GEORGE N. BENNETT*

Department of Biochemistry and Cell Biology, Rice University, Houston, Texas 77251

Received 10 June 1994/Accepted 8 September 1994

For Escherichia coli, there have been more and more examples illustrating that the a subunit of RNApolymerase is directly involved in the activation of gene transcription by interaction with activator proteins.Because of the vital function of the a subunit in cell growth, only a limited number of mutations in itsstructural gene, rpoA, have been isolated. We obtained a number of these mutants and examined the effects ofthese mutations on the acid induction of adi and cad gene expression. Several mutations caused a smallreduction in adi promoter activity at inducing pH. One mutation, rpoA341, essentially eliminated adi promoteractivity, while it had little effect on the cad promoter. During the course of a separate study, we isolated aplasmid that enhanced adi expression. Further characterization of this plasmid showed that it contained cysB,the structural gene for the positive regulator for most cys operon genes. Introduction of a cysB mutation intoan adi::lac fusion strain and 0-galactosidase assay studies of the resultant adi::ac cysB mutant established thata wild-type cysB gene was required for efficient acid induction of adi expression. These results suggest that apossible interaction between CysB and the a subunit of RNA polymerase is involved in activation of aditranscription.

The RNA polymerase holoenzyme of Escherichia coli iscomposed of core enzyme, a2pp ', and one of several kinds ofsigma (v) subunits. The core enzyme, once assembled from a,, and 1' subunits, has the ability to catalyze DNA-dependentRNA synthesis, but the sigma subunit is essential for sequence-specific interaction with and transcription from promoters (13,16). Extensive studies have been conducted to localize func-tional domains within each subunit in order to better under-stand the contribution of each subunit during the transcriptionprocesses at specific promoters. The a subunit is a peptide of329 amino acid residues encoded by the rpoA gene at 73 min onthe E. coli chromosome. Dimerization of two a subunits is thefirst step in the core enzyme assembly process, and the dimerprobably interacts with both large subunits (1 and ,B') duringthis process (13). The assembly function of the a subunit canbe provided by the N-terminal two-thirds of the protein (235residues at most), but the C-terminal one-third is required forsome functions essential for normal cell growth (8, 11, 12, 27).A number of mutants with mutations in the rpoA gene have

been isolated and characterized. A common feature amongthem is that they all showed impairment in gene expressionsystems that required a regulon- or operon-specific positiveregulator for transcription initiation, although the effects of themutations were relatively allele specific (reviewed in references15 and 27). In vitro transcription studies using three positiveregulators (cyclic AMP receptor protein [CRP], OmpR, andPhoB) and reconstituted RNA polymerase holoenzyme com-posed of C-terminally truncated a subunits and normal , 13',and a70 subunits revealed that the C-terminal one-third of thea subunit was essential for transcriptional activation on pro-moters that required activator-binding sites upstream of the-35 region (reviewed in reference 14). A contact site on the a

* Corresponding author. Mailing address: Department of Biochem-istry and Cell Biology, Rice University, P.O. Box 1892, Houston, TX77251. Phone: (713) 527-4920. Fax: (713) 285-5154. Electronic mailaddress: gbennett@bioc.rice.edu.

subunit has thus been proposed for interaction with each of theproteins that function to activate transcription through bindingto sites upstream of -35. Each contact site was proposed tocontain three to five amino acid residues within the C-terminalsegment of the a subunit (14). For those promoters that hadactivator-binding sites overlapping the -35 region, at least twoactivators (CRP and PhoB) have been suggested to makecontact with the -35 recognition domain (region 4.2) of thec70 subunit to stimulate the activity of the RNA polymerase(15).We were interested in studying the effect of rpoA alleles on

the expression of acid-induced genes (21, 32). In particular,information on the activator of the cad operon, CadC (35), andits apparent site of action on the cad promoter (21) led us toseek evidence for rpoA-activator interaction of the type de-scribed above. At the same time, studies of the effects of rpoAmutations were also conducted with the acid-induced adi gene(32). Several rpoA mutant strains defective in certain activator-regulated systems were investigated. While these rpoA muta-tions had little effect on cad expression, the introduction ofrpoA341 essentially eliminated the acid induction of the adipromoter. This result suggests that an activation of RNApolymerase through the a subunit may be involved in theregulation of adi expression.

Since one of the phenotypes conferred by this rpoA mutationis cys auxotrophy (26), our interest in this feature was en-hanced when we observed that plasmids bearing cysB couldincrease adi expression. CysB is a member of the LysR familyof regulatory proteins and is the positive regulator for mostgenes involved in the cysteine biosynthesis pathway (reviewedin reference 18). It binds upstream of the -35 region of itspositively regulated promoters and facilitates the formation ofan active transcription initiation complex in the presence of aninducer molecule, 0- or N-acetylserine. Models for CysBinteraction at various promoters have been discussed (10). Wereport here the finding that a wild-type cysB gene is requiredfor efficient acid induction of adi expression.

on February 4, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

7018 SHI AND BENNEIT

TABLE 1. Bacterial strains and plasmids

Strain or plasmid Genotypea reference

Bacterial strainsDZ3-1 Hfr P02A metB gltS0 lacYphs (L271E) zhc::TnlO 26EC1173 araDl39 A(lac)U169 strA relA thi trp cysB88::Mud(Apr lac) 2JMS4540 MC4100 araD+ 4'(ompF'-lacZ+)16-13 zhc-3::TnlO 31JMS4542 JMS4540 rpoA85 (P323L) 31JMS4543 JMS4540 rpoA50 (L28F/P240S) 31JMS4544 JMS4540 rpoA52 (P323S) 31JMS4545 JMS4540 rpoA53 (G3S) 31JMS4546 JMS4540 rpoA54 (P322S) 31MH513 MC4100 araD+ 4(ompF'-lacZ+)16-13 31MDA4762 cysB trpB::TnlO E. MurgolaMDA7K GNB7145K cysB trpB::TnlO This studyGNB7145K MC4100 adi::MudI 1734(Kmr lac) 1GNB725 GNB7145K hns-25::Tn5(Cmr) 29

PlasmidspMAN225 rpoA+ controlled by lpp-lac-promoter-operator 20pJS115 pBR322plac rpoA+ 31pLAW rpoA+ under the control of lpp-lac, Apr 37pLAMC3 pLAW rpoA123 (A802G/N268D) 37pLAMC5 pLAW rpoA125 (G794A/R265H and A890G/K297R) 37pLAMC7 pLAW rpoA127 (T890C/L279P) 37pLAMC9 pLAW rpoA129 (C793T/R265C) 37pMAHOl cysBc301 on -2.8-kb EcoRI-SalI insert in pBR322, Apr 9pMAH2 cysBc302 on -5-kb EcoRI insert in pBR322, Apr Tcr 9pMAH3 cysBc303 on -5-kb EcoRI insert in pBR322, Apr Tcr 9plO6T9 pBR322 cysB+ topA+ Apr This studyp106T18 Derivative of plO6T9, cysB+ Apr This studyplO6RI 3-kb EcoRI fragment from plO6T9 subcloned into the EcoRI site in pEMBL8+, cysB+ Apr This study

a For some of the rpoA mutants (JMS strains) and rpoA plasmids (pLAMC plasmids), the changes in the nucleotide sequence of the rpoA structural gene and in theamino acid sequence of the a subunit of RNA polymerase caused by the mutations are indicated in parentheses.

MATERIALS AND METHODS

Materials. Restriction enzymes were purchased from Pro-mega Biotech and were used under the conditions recom-mended by the manufacturer. A Prep-A-Gene DNA purifica-tion kit was purchased from Bio-Rad. Most chemicals were

purchased from Sigma Chemical Company. Bacterial growthmedia were from Difco, and buffers were from ResearchOrganics. Plasmid DNA preparation kits were from Qiagen.The gene-mapping membrane containing an aligned E. colilambda library was purchased from Takara Shuzo Co. Ltd.

Bacterial strains, plasmids, growth media, and buffers. Thebacterial strains and plasmids used in this study are listed inTable 1. P1 transduction was carried out as described previ-ously (30). L broth contains (in 1 liter) 10 g of tryptone, 5 g ofyeast extract, 10 g of sodium chloride, and 30 mg of cysteine.Buffered modified Falkow arginine medium contains (in 1liter) 5 g of Bacto-Peptone, 3 g of yeast extract, 5 g ofarginine-HCl, 10 g of glucose, and 19.5 g of MES [2-(N-morpholino)ethanesulfonic acid] for pH 5.5 medium or 23.8 gof HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonicacid) for pH 8.0 medium. The pHs of the media were adjustedwith concentrated NaOH before autoclaving.Enzyme assays. 1-Galactosidase assays in modified Falkow

medium under partial anaerobic conditions were as describedpreviously (29).DNA techniques. Large-scale plasmid DNA purification was

done by using the Qiagen midi-plasmid preparation kit accord-ing to the manufacturer's specifications. Plasmid DNA was

transformed into host cells by the method of Chung and Miller(3). Restriction fragments used for subcloning were purified

from low-melting-point agarose gels (NuSieve GTG fromFMC) by using the Prep-A-Gene kit. Ligation reactions werecarried out under standard conditions (28).

RESULTSEffects of rpoA mutations on expression of acid-induced

amino acid decarboxylase genes. We obtained several drugmarker-linked rpoA alleles that had one or two amino acidchanges in the a subunit (Table 1) and examined the expres-sion of arginine decarboxylase or lysine decarboxylase in thosestrains (1). No substantial effect was observed in the case oflysine decarboxylase, even with rpoA341, an allele which dra-matically affected adi. Since the effects on adi were readilyobserved, the analysis was pursued further and is specificallydescribed here.The rpoA allele and the linked drug marker (zhc-3::TnlO-Tc`)

were transduced into the adi::lac fusion strain GNB7145K andone of its hns mutants, GNB725 (adi::lac hns-25::Tn5_Cmr). hnsmutants were used because the ,B-galactosidase values werehigher and a difference might be more easily observed, espe-cially at the noninducing pH (pH 8.0). Since the drug markerswere only close to, not 100% linked to, the rpoA alleles, wechose more than eight colonies from each transduction andassayed the P-galactosidase activities expressed from theadi::lac fusion. In the case of rpoA341 from DZ3-1, the Cym-phenotype, i.e., the requirement for methionine or cysteine forgrowth, was used as a marker for screening rpoA341 transduc-tants (26). In all but one case (rpoA52 zhc-3::TnlO intoGNB7145K), one or more transductants were found to havedecreased ,B-galactosidase activities at pH 5.5 compared with

J. BACrERIOL.

on February 4, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 176, 1994 rpoA AND cysB IN adi REGULATION 7019

TABLE 2. Effects of some rpoA mutant alleles on adipromoter activity'

li-Galactosidase activity' in strain:rpoA Location' ad::lac hns-25allele adi::lac (pH 5.5)

pH 8.0 pH 5.5

rpoA85 323 230 ± 10 28 ± 2 658 ± 20rpoASO 28, 240 127 ± 21 15 ± 2 489 ± 9rpoA52 323 31 ± 2 633 ± 43rpoA53 3 62 ± 13 17 ± 1 320 ± 41rpoA54 322 83 ± 8 17 ± 4 440 ± 18rpoA341 271 1.5 ± 0.2 1.6 ± 0.2 1.9 ± 0.2rpoA+ 315 ± 9 79 ± 10 1,216 ± 65

a The various rpoA alleles were introduced into the indicated strains (adi::lacor adi::lac hns-25) by Pl transduction. The Il-galactosidase activities of at leastthree independent colonies were assayed for each transductant.b Amino acid position(s) of the change(s) in each mutant ax subunit.

The values shown are averages ± standard deviations. The units are asdefined by Miller (22).

the wild type. Since all of the rpoA alleles in JMS strains wereinitially isolated for their abilities to decrease the P-galactosi-dase activity produced from the ompF'-lacZ+ operon fusion instrain MH513 (Table 1) (31), we transduced rpoAzhc-3::TnlO-Tcr from the down-regulating GNB7145K trans-ductants back into MH513 and assayed the 3-galactosidaseactivities of the transductants by the procedures describedpreviously (31). This backcross showed that these transduc-tants exhibited the expected 65% linkage between the TnlOmarker and the mutant rpoA phenotype (i.e., reduced ompFexpression). As a control, backcrosses of some GNB7145Kzhc-3::TnlO strains which exhibited the same f-galactosidaseactivities as the parent GNB7145K all resulted in Tcr trans-ductants with 100% linkage between the wild-type rpoA phe-notype and the TnlO marker, indicating that the adi-down-regulating phenotype is associated with the rpoA mutations.Table 2 shows the 13-galactosidase assay values for selectedGNB7145K rpoA mutant strains grown without aeration at pH5.5. To examine the effect more easily at pH 8.0, an hns strainderepressed for adi was also investigated. The same cycles ofP1 transduction, ,B-galactosidase assays, and Pl backcrosseswere also performed with an adi::lac hns mutant, GNB725(Table 1), and the assay values for some GNB725 rpoA mutanttransductants at both pH 5.5 and pH 8.0 are shown in Table 2.From the assay values shown in Table 2, we can conclude

that each rpoA mutation caused a reduction of at least twofoldin the expression of 3-galactosidase from the adi::lac fusion. Inthe case of rpoA341, when it was introduced into eitherGNB7145K or GNB725, essentially no 3-galactosidase expres-sion (<2 Miller units) could be detected from the adi::lacfusion at either pH. This result was specific for adi sincecorresponding cad::lac and cad::lac hns fusion strains analyzedin the same fashion showed no effect due to rpoA341. Thedramatic reduction of adi expression suggested an importantrole of lysine residue 271 in the ability of adi to be induced byacid. The effects of the other rpoA alleles on adi may bethrough a different mechanism.Mapping of the contact site on the a subunit for the putative

activator of adi. Two rpoA+ plasmids, pMAN225, in whichrpoA+ is under the control of the lpp-lac-promoter-operator(Table 1) (20), and pJS115, in which rpoA + is under the controlof plac (Table 1) (31) were used. The expression of thewild-type rpoA gene in both plasmids can be stimulated by thepresence of the inducer IPTG (isopropyl-13-D-thiogalactopyr-anoside). When these plasmids were introduced into

TABLE 3. Effects of introduction of plasmid-encoded rpoA mutantalleles on ,-galactosidase expression from adi::lac rpoA341a

,-Galactosidase activityc inadi::lac rpoA34J cells:Plasmid Alterations (nt/aa)b

With WithoutIPTG IPTG

None 3.5 ± 0.2 6.7 ± 1.2rpoA+ 45 ± 6 104 ± 7rpoA123 A802G/N268D 3.8 ± 0.3 6 ± 0.1rpoA125 G794A/R265H, 45 ± 11 107 ± 5

A890G/K297RrpoA127 T809C/L270P 3.8 ± 0.2 7 ± 0.6rpoA129 C793T/R265C 4.0 ± 0.2 7 ± 0.5

a Plasmids carrying different rpoA alleles under the control of the lpp-lac-promoter-operator system (Table 1) were transformed into adi::lac rpoA341cells, and levels of ,B-galactosidase expression from the adi::lac fusion wereassayed after growth in modified Falkow medium (pH 5.5) as described inMaterials and Methods. When used, IPTG was added to modified Falkowmedium at 40 ,ug/ml to fully induce the expression of the rpoA alleles.

b Changes in the nucleotide sequence of the rpoA coding region (nt) and thecorresponding changes in the peptide sequence of the a subunit of RNApolymerase (aa) for the various rpoA alleles, as described by Zou et al. (37).c Shown are averages of three independent assays with standard deviations.The units are as defined by Miller (22).

GNB7145K rpoA341, they were able to restore to some extentadi promoter activity at the inducing pH (pH 5.5) as measuredby 3-galactosidase activities expressed from the adi::lac fusion,especially in the presence of IPTG. A typical value is repre-sented by that for rpoA+ in Table 3. The 3-galactosidase valuesfor the rpoA+-complemented GNB7145K rpoA341 strain in-creased only to one-third to one-half of that of the wild-typestrain, GNB7145K, because an excess amount of a subunitcould destroy the stoichiometric balance among the differentsubunits of the RNA polymerase and interfere with the normalfunction of this enzyme. This reduced expression has beenreported earlier for promoters whose activation requires apositive regulator contacting the a subunit for its function (31).A group of plasmids containing an rpoA gene with pointmutations causing amino acid changes in the region fromposition 265 to 270 of the a subunit had been characterized fortheir inability to transcribe a number of CRP-activated genes(37). Contact site I for CRP on the a subunit was thereforemapped to the region containing at least six amino acidresidues from position 265 to 270. Akira Ishihama kindlyprovided these plasmids containing point-mutated rpoA genes(Table 1). When transformed into the adi::lac rpoA341 strain,none of the single point mutant alleles could complement thephenotype conferred by the rpoA341 mutation on adi expres-sion (the double point mutant allele rpoA125 behaved like awild-type rpoA allele) (Table 3), suggesting that all produced asubunits incapable of proper interaction with a positive regu-lator required for adi activation. In rpoA125, the secondmutation may serve as a suppressor for the mutation at residue265. In summary, this result, as well as those shown in Table 2,indicated a possible contact site on the a subunit that wasessential for adi activation within a region containing at leastseven amino acid residues from position 265 to 271, whichoverlaps contact site I for CRP.

Defining a role for cysB in activation of expression of adi. Ifan activator interacting with the a subunit of RNA polymeraseis required for adi induction, what are possible candidates forthis activator? Some insight might be gained by considerationof the other phenotypes conferred by the rpoA341 allele (26).One of these phenotypes is Cym-, which could be due to thelack of ability to express genes of the cys regulon, a property

on February 4, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

7020 SHI AND BENNETT

1 kb

EcoRI PstIPvUI BamHI

EcoRI |I l

topA

BamHI

cysB

p1O6T18 (pBR322)

p1O6RI (pEMBL8+)FIG. 1. Subcloning of plasmid plO6T9. At top is the restriction enzyme map of the insert on plO6T9, constructed from our restriction enzyme

mapping data and the published nucleotide sequences of topA (34) and cysB (24). The locations and directions of transcription of the topA andcysB genes are indicated. The restriction fragments subcloned are shown as thin lines below the p1O6T9 restriction map. Vectors used are shownin parentheses. Dashes at the ends of some lines indicate vector segments that were also brought into the new constructs.

also conferred by mutations in cysB, a known positive activator.Therefore, an examination of adi induction in cys mutantstrains was undertaken. Another approach to identifying an

activator is to seek genes that can lead to increased expressionof adi when they are overexpressed from high-copy-numberplasmids. This expected phenotype of an activator gene was

displayed by cadC, the activator required for cad operonexpression (21, 35). Results from this approach also implicatedcysB as being involved in the activation of adi. The results ofthese two lines of research are described separately below.

In an effort to identify a gene activating adi, plasmids froman E. coli library (29) that could elevate the expression of3-galactosidase in an adi::lac fusion strain were identified. A

plasmid, p106T9, was able to increase expression from the adipromoter to a level higher than that of the wild-type strain(GNB7145K) (see Table 6). When a radioactive probe pre-pared from p106T9 was used to screen a gene-mappingmembrane containing an aligned E. coli lambda library(Takara Shuzo Co. Ltd.), by the procedure described previ-ously (23), two phages, 4F1 (serial no. 253) and 13F9 (serial no.

254), showed strong positive signals, indicating that they bothcontained the insert sequence on plasmid plO6T9. By compar-ing the preliminary restriction enzyme map of p106T9 with therestriction map of the overlapping region of the fragmentscarried by these two phage clones, which mapped at 28.45 to28.75 min on the E. coli chromosome, we deduced that thisplasmid carried both cysB and topA genes (Fig. 1). Afterdigestion with PvuII, the larger fragment (7.3 kb) of p106T9was recircularized to yield a subclone, p106T18, which retainedthe ability to elevate expression from adi::lac fusions (Table 4).A ca. 3-kb EcoRI fragment from p106T18, containing a smallC-terminal region of topA, the entire cysB gene, and a fewhundred base pairs of DNA downstream of cysB, was sub-cloned into pEMBL8+ to yield plasmid plO6RI (Table 1 andFig. 1). This plasmid also retained the elevated-expressionphenotype of the previous two larger clones, plO6T9 andp106T18. Cells bearing plO6RI grew very slowly, especially atpH 5.5, perhaps because the high copy number allowedexcessive expression of a product (likely CysB) which impededgrowth.There are three BamHI sites in plO6RI, one of which lies

near the 3' end of the cysB coding sequence (Fig. 1). Digestion

of p106RI with BamHI cleaved the insert-vector junction andthe site within cysB, yielding two fragments of about 1 kb.Cloning of each BamHI fragment into the BamHI site ofpEMBL8+ resulted in two constructs, neither of which had theability to increase expression in the adi::lac fusion strain. Thissuggested that neither fragment was sufficient for the pheno-type. We thus concluded that the phenotype of the p106plasmid series was caused by overproduction of CysB proteinfrom the cysB gene carried on the plasmids.To test further the properties of activation by CysB, cysB+

plasmids plO6T9, p106T18, and p106RI were transformed intothe adi::lac strain GNB7145K and its hns derivative GNB725.In each strain, the plasmids led to increased adi promoteractivity (two- to threefold increase) (Table 4), suggesting ageneral adi-activating role of excess CysB protein. This acti-vating effect is gene specific because these plasmids did notshow any influence on the gene expression of another acid-inducible decarboxylase, the lysine decarboxylase encoded bycadA (data not shown).

Requirement for a wild-type cysB gene in the acid inductionof adi expression. The cysB mutation from a trpB::TnlO cysBstrain, MDA4762 (Table 1), in which cysB is about 60% linkedto the Tn1O-Tcr marker in trpB, was introduced intoGNB7145K by P1 transduction. Tcr transductants unable to

TABLE 4. Activating effects of p106 plasmids in GNB7145Kand GNB725

,3-Galactosidase activity" in strain:

Plasmida GNB7145K (adi::lac) GNB725 (adi::lac hns-25)

pH 8.0 pH 5.5 pH 8.0 pH 5.5

None 5.0 ± 0.26 546 ± 58 166 ± 31 1,411 ± 162pBR322 4.6 ± 0.35 680 ± 31 103 ± 22 1,106 ± 103plO6T9 18 ± 3.0 1603 ± 157 227 ± 21 1,928 ± 138p106T18 19 ± 2.6 1786 ± 14 255 ± 14 2,400 167plO6RI 26 ± 2.6 1,060 ± 72 222 ± 18 3,427 + 100

a Competent cells alone (None) and cells transformed with the vector,pBR322, were included as controls.

' Presented are average values of three independent assays with standarddeviations. The units are as defined by Miller (22).

I I I i i i I -9 i -2

J. BACTERIOL.

on February 4, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

rpoA AND cysB IN adi REGULATION 7021

TABLE 5. Complementation of cysB by cysB+ and cysBc plasmidsa

f3-Galactosidase activity in strain:

Plasmid adi::lac cysB adi::lac hns-25 cysB

pH 8.0 pH 5.5 pH 8.0 pH 5.5

None 1.0 ± 0.6 7 ± 2 2± 0.1 27 ± 8pBR322 10±2-±-p106T18 907 ± 325 37 ± 2 1,330 ± 60

(cysB+)pMAHO1 21 ± 4 935 ± 19 52 ± 1 1,380 ± 68

(cysBc301)pMAH2 30 ± 3 1,302 ± 186 49 ± 5 2,240 ± 196

(cysBc302)pMAH3 23 ± 4 1,165 ± 157 41 ± 3 1,430 ± 186

(cysBc303)a The TnlO-linked cysB mutation was P1 transduced into GNB7145K (adi::lac)

and GNB725 (adi::lac hns-25) to obtain the two cysB mutant strains MDA7K andMDA725, respectively.

b The values are averages of three independent assays with standard devia-tions, and the units are as defined by Miller (22). Dashes indicate that the datawere not determined. The values for the adi::lac fusion strain (GNB7145K) were5.4 ± 0.9 at pH 8.0 and 520 ± 61 at pH 5.5. The values for the adi::lac hns-25strain (the cysB+ counterpart of MDA725) were 65 ± 5 at pH 8.0 and 1,580 ±55 at pH 5.5.

grow on minimal plates supplemented with tryptophan butwithout cysteine were the desired adi::lac cysB mutants. The3-galactosidase activity of one such transductant, MDA7K,

was assayed (Table 5). The presence of the cysB mutationcaused the value at pH 8.0 to decrease about 5-fold and causedthe value at the inducing pH, pH 5.5, to be reduced more than30-fold. There was still a small acidic pH induction eventhough the absolute enzymatic activities were very low at bothpHs. When the cysB+ containing plasmids, p1O6T9 andp106T18, were transformed into the adi::lac cysB mutantstrain, they restored ,-galactosidase activity at the inducingpH, pH 5.5, to a level higher than that of the parental strain(Table 5). Thus, the excess CysB protein produced from theplasmids has a general adi-activating phenotype whether thechromosomal copy of cysB is wild type or mutated (compareTables 4 and 5).To better examine the effect of the cysB deficiency at high

pH, a derepressed mutant, carrying hns-25, was employed,since it normally produces high levels of P-galactosidase at pH8. Essentially the same trend was observed when a cysBmutation was introduced into an adi::lac hns mutant, GNB725.In the resultant transductant, MDA725 (adi::lac hns-25 cysB),the P-galactosidase activities were at least 50-fold lower thanthose of GNB725 at both pH 8.0 and 5.5, but there was stillabout a 10-fold induction at pH 5.5 versus pH 8.0. The enzymeactivity of the hns cysB strain at pH 5.5 was ca. threefold higherthan that of the cysB strain (Table 5), suggesting that changesin DNA topology in an hns mutant retained some effect on adiexpression (at pH 5.5) in the absence of a functional CysBprotein. In the presence of p106T18, the enzymatic activitieswere increased to about the same level as in GNB725 at pH 5.5(>50-fold increase) and to one-half to one-third of the value inGNB725 at pH 8.0 (-20-fold increase) (Table 5).

Since the function of CysB as an activator for cys operongenes is antagonized by sulfur compounds (including cysteine)which exert their effects by competing with the inducers 0- andN-acetylserine (18) and since such sulfur compounds arepresent in the rich medium (modified Falkow) that we used for,B-galactosidase assays, we tested the effects of three plasmidscarrying cysB constitutive mutations (cysBc301 to cysBc3o3) on

TABLE 6. 3-Galactosidase assays with cysB::lac strain EC1173a

,B-Galactosidase activity'Plasmid

o/n LB pH 8.0 pH 5.5

None 867 ± 28 509 ± 21 368 ± 17pBR322 1,049 ± 38 621 ± 24 424 ± 11p106T18 (cysB+) 545 ± 62 330 ± 50 308 ± 52

a The ,B-galactosidase activities were assayed after the cells were grown in Lbroth overnight (o/n LB) or to late log phase (optical density at 600 nm of 0.4 to0.6) in buffered modified Falkow medium, as described in Materials andMethods.bShown are averages of three or four independent assays with standard

deviations. The units are as defined by Miller (22).

adi expression levels in strains MDA7K (adi::lac cysB) andMDA725 (adi::lac hns-25 cysB) (Table 5). The phenotypecaused by these constitutive mutations is constitutive expres-sion of the cys regulon (a Cys+ phenotype) regardless ofacetylserine and sulfur source (9), suggesting that the alteredCysBc proteins mimic the active form of CysB bound by aninducer. We reasoned that if the function of CysB in activatingadi expression was induced by the same inducer molecule(s)acting in a pH-sensitive fashion, then these constitutive muta-tions might cause adi to be expressed at high levels regardlessof the medium pH. If no inducer or a different inducer wasused by CysB in regulating adi expression, these cysBc mutantswould behave the same as a wild-type cysB gene. The data inTable 5 indicate that the cysBc genes showed the samemagnitude of effect as a cysB+ plasmid under both conditions,implying that CysB protein may exert its activating function inthe cys regulon and in the adi gene system via differentmechanisms.

Overall, these experiments showed that a functional cysBgene was required for efficient acid induction of adi promoteractivity, whether the background was hns+ or hns mutant. Wethus consider CysB to be a candidate activator protein forinduction of adi expression at optimum inducing conditions.

cysB expression in response to growth pH. The amount oractivity of CysB might be altered by culture pH. To investigatethese possibilities, we obtained a cysB::lac fusion strain,EC1173 (Table 1), and used it to study cysB gene expressionlevels in response to medium pH in rich broth. As can be seenfrom Table 6, the ,B-galactosidase values of stationary-phasecells were approximately twofold greater than those at logphase, and in log phase the assay values were approximatelythe same at the two tested pHs, suggesting that medium pHdoes not affect the level of cysB expression. When the cysB+- orcysBc-containing plasmids were transformed into the cysB::lacfusion strain, the P-galactosidase values were essentially un-changed. This result indicates that the presence of excess CysBprotein had little repressive effect on cysB expression underthese conditions. In a previous assay conducted with minimalmedium supplemented with L-cystine or L-djenkolic acid, thepresence of excess CysB production resulted in a ca. 10-foldreduction in the P-galactosidase activity from a chromosomalcysB::lac fusion (17).

DISCUSSION

Several examples of negative effects of rpoA point mutationson activation of gene systems have been reported. This cumu-lative information supports the idea that a group of (type I)activators facilitate the transcription activity of RNA poly-merase by directly interacting with the a subunit. Specificexamples include the following: rpoA109, mapped at residue

VOL. 176, 1994

on February 4, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

7022 SHI AND BENNETI

290, which impairs the function of the phage late gene positiveregulator encoded by the P2 ogr or P4 8 gene (7); rpoA8, -151,-153, -154, and -155, which reduce the anaerobic induction ofthe genes requiring the positive regulator OxrA (FNR) inSalmonella typhimurium (19); and three groups of rpoA muta-tions, (i) rpoA77 (szr), (ii) rpoA50, -52, -53, and -54 (dip), and(iii) sezl8, -85, -44, and -611, all of which show changes intranscriptional activation of porin gene expression involvingthe EnvZ-OmpR two-component regulatory system (4, 20, 31).Studies on the involvement of the RNA polymerase a subunitin the activation of transcription from several promoters byCRP have revealed a possible contact site for interaction withCRP on the cx subunit at residues 265 to 270 (37). Thisinteraction was found to be essential for CRP activation of(type I) promoters that have CRP-binding sites upstream of-35 regions.We have shown that several rpoA mutations reduced 3-ga-

lactosidase expression from an adi::lac operon fusion, by atleast twofold. One mutation, rpoA341, causing a single aminoacid change at residue 271, resulted in an RNA polymerasethat could support bacterial growth but essentially eliminatedadi expression at inducing pH (pH 5.5). We mapped thepossible contact site on the a subunit for a potential adi-activating protein to be in a region partially overlappingcontact site I for CRP, including amino acid residues 265 to271. Since there seems to be an absolute requirement for CysBin order to achieve efficient induction of adi expression, weconsider CysB a reasonable candidate for the positive regula-tor of adi and propose that adi be added to the list of genesinfluenced by the rpoA341 mutation, joining the melAB operon(MeiR as the positive regulator), the ara regulon (AraC as theactivator), the cys operon (CysB as the activator), and thegenes involved in phage X lysogenization requiring activatorproteins CI and CII (5, 33, 36).

Several results relevant to the specificity or mechanism ofCysB function in this system are considered. While inhibitorsof DNA gyrase have been shown to reduce adi expression (32),DNA supercoiling is unaffected in cysB mutants (25). However,expression of cysB was decreased upon inhibition of DNAgyrase (2, 6), and this reduced level of CysB could lead tolower-level expression of adi. Such a mechanism would beconsistent with the observed effects of the gyrase inhibitors onadi expression. Whether the effect on cysB is sufficient tocompletely explain the magnitude of the reduction of adiexpression observed with novobiocin (32) is unclear. Wetransformed strains GNB7145K rpoA341 and GNB7145K hnsrpoA341 with cysB+ plasmids and found that the excess amountof CysB protein produced was unable to rescue adi expressionfrom the effect of the rpoA341 mutation (data not shown). Ithad been previously reported that rpoA341 did not affecttranscription of cysB (5), so the rpoA341 mutation apparentlydoes not act by merely lowering CysB levels. This result isconsistent with our proposition that the potential activatorprotein, CysB, directly interacts with the ot subunit to stimulatethe transcriptional activity of the RNA polymerase. Thus, withthe loss of the RNA polymerase contact site, the activatorprotein, even at high levels that are able to "overstimulate" thewild-type RNA polymerase, was not able to interact with therpoA341 mutant RNA polymerase to exert its activation func-tion. Also in support of this proposition is a previouslyreported phenotype of the rpoA341 mutant, the inability tosynthesize cysteine, indicating a defect in the cys regulonpositively regulated by CysB (5). It is possible, therefore, thatCysB activates both cys genes and adi by interacting with thesame contact site on the a subunit. The rpoA contact sitedefined here with regard to adi expression may represent that

contacted by CysB. Studies of the expression of various cysB-regulated genes showed that only a cysA::lac fusion wasaffected by rpoA341 and not cysK::lac or cysI::lac fusions (5).Therefore, in terms of CysB interactions, the most relevantexample would be those at the cysP promoter.

It is also possible that a protein product or small moleculeprovided by a CysB-regulated gene is necessary for the pro-duction of a signal or effector molecule used by some otherDNA-binding activator molecule which acts on adi. O-Acetyl-serine and N-acetylserine have been found to be inducermolecules whose binding to CysB is required for the activationof CysB-regulated cys promoters. This function is antagonizedby sulfur compounds (18). Perhaps a small signal molecule thatcould sense the changes in the environmental pH and couldmodulate the function of CysB according to pH exists. Theidentity of such an inducer molecule, if it exists, is presentlyunknown. The facts that mutations in cysE, which encodes anenzyme (serine acetyltransferase) required for the synthesis ofO-acetylserine, have no effect on adi expression (data notshown) and that constitutive cysB alleles had the same magni-tude of adi-activating ability as the wild-type cysB allele in richmedium (Table 5) suggest that these effectors may not beinvolved in the activation of adi expression. Since the expres-sion of the cysB gene does not change with changes in growthpH (Table 6), the adi-activating activity is likely the featureregulated as a result of pH changes, either by a small signalmolecule or by another protein(s) involved in forming afunctional transcriptional complex. Thus, the mechanism in-volved here appears to be different from that involved in cysregulon regulation.

Other physiological effects have been observed in cysBmutants. Increased resistance of cysB strains to novobiocin(25) and increased resistance to Cd2+ (6) suggest a role in theexpression of other genes. Perhaps the mechanism of theseeffects has features in common with the involvement of CysB inadi regulation. Experiments to elucidate the role of cysB inother systems would expand our understanding of the overallphysiological significance of this important regulator.

ACKNOWLEDGMENTS

We thank Bernardina Waasdorp for isolation of and earlier exper-iments with the GNB7106 mutant, and we thank David Mendoza forassistance in isolation and preliminary characterization of the comple-menting cysB+ plasmids. We also thank Thomas Silhavy and AkiraIshihama for providing the mutant rpoA strains and plasmids, respec-tively.

This work was supported by National Science Foundation grantBCS-9315797 and by the Texas Advanced Technology Program (grant003604-035).

REFERENCES1. Auger, E. A., K. E. Redding, T. Plumb, L. C. Childs, S.-Y. Meng,

and G. N. Bennett. 1989. Construction of lac fusions to theinducible arginine- and lysine decarboxylase genes of Escherichiacoli K-12. Mol. Microbiol. 3:609-620.

2. Bielinksa, A., and D. Hulanicka. 1986. Effect of DNA gyraseinhibitors and urea on the expression of cysB, the regulatory geneof the cysteine regulon. J. Gen. Microbiol. 132:2571-2576.

3. Chung, C. T., and R. H. Miller. 1988. A rapid and convenientmethod for the preparation and storage of competent bacterialcells. Nucleic Acids Res. 16:3580.

4. Garrett, S., and T. J. Silhavy. 1987. Isolation of mutations in the aoperon of Escherichia coli that suppress the transcriptional defectconferred by a mutation in the porin regulatory gene envZ. J.Bacteriol. 169:1379-1385.

5. Giffard, P. M., and I. R Booth. 1988. The rpoA341 allele ofEscherichia coli specifically impairs the transcription of a group ofpositively regulated operons. Mol. Gen. Genet. 214:148-152.

J. BACTERIOL.

on February 4, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

rpoA AND cysB IN adi REGULATION 7023

6. Goel, R, A. Motha, R. G. Martin, and J. L. Rosner. 1992. Effectsof salicylate, supercoiling and cysB mutations on Cd"+ sensitivityin Escherichia coli, abstr. H-21. Abstr. 92nd Gen. Meet. Am. Soc.Microbiol. 1992. American Society for Microbiology, Washington,D.C.

7. Hailing, C., M. G. Sunshine, K. B. Lane, E. W. Six, and RCalendar. 1990. A mutation of the transactivation gene of satellitebacteriophage P4 that suppresses the rpoA109 mutation of Esch-erichia coli. J. Bacteriol. 172:3541-3548.

8. Hayward, R S., K. Igarashi, and A. Ishihama. 1991. Functionalspecialization within the a subunit of Escherichia coli RNApolymerase. J. Mol. Biol. 221:23-29.

9. Hryniewicz, M., A. Palucha, and M. D. Hulanicka. 1988. Construc-tion of cys::lac gene fusions in Eschenchia coli and their use in theisolation of constitutive cysBe mutants. J. Gen. Microbiol. 134:763-769.

10. Hryniewicz, M. M., and N. M. Kredich. 1994. Stoichiometry ofbinding of CysB to the cysJIH, cysK, and cysP promoter regions ofSalmonella typhimunum. J. Bacteriol. 176:3673-3682.

11. Igarashi, K., N. Fujita, and A. Ishihama. 1990. Sequence analysisof two temperature-sensitive mutations in the alpha subunit gene(rpoA) of Escherichia coli RNA polymerase. Nucleic Acids Res.18:5945-5948.

12. Igarashi, K., N. Fujita, and A. Ishihama. 1991. Identification of asubunit assembly domain in the alpha subunit of Eschenchia coliRNA polymerase. J. Mol. Biol. 218:1-6.

13. Ishihama, A. 1990. Molecular assembly and functional modulationof Eschenchia coli RNA polymerase. Adv. Biophys. 26:19-31.

14. Ishihama, A. 1992. Role of the RNA polymerase ot subunit intranscription activation. Mol. Microbiol. 6:3283-3288.

15. Ishihama, A. 1993. Protein-protein communication within thetranscription apparatus. J. Bacteriol. 175:2483-2489.

16. Ishihama, A., N. Fujita, K. Igarashi, R Ueshima, M. Nakayama,and Y. Yamazaki. 1991. Molecular mechanisms of transcriptionregulation: promoter selectivity of RNA polymerase. Recent Adv.Biochem. 17:179-185.

17. Jagura-Burdzy, G., and D. Hulanicka. 1981. Use of gene fusions tostudy expression of cysB, the regulation gene of the cysteineregulon. J. Bacteriol. 147:744-751.

18. Kredich, N. M. 1992. The molecular basis for positive regulation ofcys promoters in Salmonella typhimurium and Escherichia coli.Mol. Microbiol. 6:2747-2753.

19. Lombardo, M.-J., D. Bagga, and G. C. Miller. 1991. Mutations inrpoA affect expression of anaerobically regulated genes in Salmo-nella typhimurium. J. Bacteriol. 173:7511-7518.

20. Matsuyama, S., and S. Mizushima. 1987. Novel rpoA mutationthat interferes with the function of OmpR and EnvZ, positiveregulators of the ompF and ompC genes that code for outer-membrane proteins in Eschenchia coli K-12. J. Mol. Biol. 195:847-853.

21. Meng, S.-Y., and G. N. Bennett. 1992. Regulation of the Esche-richia coli cad operon: location of a site required for pH induction.

J. Bacteriol. 174:2659-2669.22. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring

Harbor Laboratory, Cold Spring Harbor, N.Y.23. Noda, A., J. B. Courtright, P. F. Denor, G. Webb, Y. Kohara, and

A. Ishihama. 1991. Rapid identification of specific genes in E. coliby hybridization to membranes containing the ordered set ofphage clones. BioTechniques 10:474-477.

24. Ostrowski, J., G. Jagura-Burdzy, and N. M. Kredich. 1987. DNAsequences of the cysB regions of Salmonella typhimurium andEscherichia coli. J. Biol. Chem. 262:5999-6005.

25. Rakonjac, J., M. Milic, and D. J. Savic. 1991. cysB and cysEmutants of Escherichia coli K12 show increased resistance tonovobiocin. Mol. Gen. Genet. 228:307-311.

26. Rowland, G. C., P. M. Giffard, and I. R. Booth. 1985. phs locus ofEscherichia coli, a mutation causing pleiotropic lesions in metab-olism, is an rpoA allele. J. Bacteriol. 164:972-975.

27. Russo, F. D., and T. J. Silhavy. 1992. Alpha: the Cinderella subunitof RNA polymerase. J. Biol. Chem. 267:14515-14518.

28. 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.

29. Shi, X., B. C. Waasdorp, and G. N. Bennett. 1993. Modulation ofacid-induced amino acid decarboxylase gene expression by hns inEscherichia coli. J. Bacteriol. 175:1182-1186.

30. Silhavy, T. J., M. L. Berman, and L. W. Enquist. 1984. Experi-ments with gene fusions. Cold Spring Harbor Laboratory, ColdSpring Harbor, N.Y.

31. Slauch, J. M., F. D. Russo, and T. J. Silhavy. 1991. Suppressormutations in rpoA suggest that OmpR controls transcription bydirect interaction with the a subunit of RNA polymerase. J.Bacteriol. 173:7501-7510.

32. Stim, K. P., and G. N. Bennett. 1993. Nucleotide sequence of theadi gene, which encodes the biogradative acid-induced argininedecarboxylase of Escherichia coli. J. Bacteriol. 175:1221-1234.

33. Thomas, M. S., and R E. Glass. 1991. Escherichia coli rpoAmutation which impairs transcription of positively regulated sys-tems. Mol. Microbiol. 5:2719-2725.

34. Tse-Dinh, Y.-C., and J. C. Wang. 1986. Complete nucleotidesequence of the topA gene encoding Escherichia coli DNA topoi-somerase I. J. Mol. Biol. 191:321-331.

35. Watson, N., D. S. Dunyak, E. L. Rosey, J. L. Slonczewski, and E. R.Olson. 1992. Identification of elements involved in transcriptionalregulation of the Escherichia coli cad operon by external pH. J.Bacteriol. 174:530-540.

36. Wegryn, G., R E. Glass, and M. S. Thomas. 1992. Involvement ofthe Eschenchia coli RNA polymerase a subunit in transcriptionalactivation by the bacteriophage lambda CI and CII proteins. Gene122:1-7.

37. Zou, C., N. Fujita, K. Igarashi, and A. Ishihama. 1992. Mappingthe cAMP receptor protein contact site on the a subunit ofEscherichia coli RNA polymerase. Mol. Microbiol. 6:2599-2605.

VOL. 176, 1994

on February 4, 2021 by guest

http://jb.asm.org/

Dow

nloaded from