JOURNAL OF BACTERIOLOGY, May 1975, p. 510-517Copyright 0 1975 American Society for Microbiology

Vol. 122, No. 2Printed in U.S.A.

Regulation of the Galactose Operon at the galOperator-Promoter Region in Escherichia coli K-12

SUI-SHENG HUA AND ALVIN MARKOVITZ*

Department of Microbiology, University of Chicago, Chicago, Illinois 60637

Received for publication 10 February 1975

The capR (lon) product controls expression of the gal operon independently ofthe galR repressor. Previously, mutations of the gal operon have been isolatedthat are semiconstitutive and alter response to the capR and/or capT product.Such mutants imply the existence of a distinct site in the operon that responds tocapR (capT) control. This mutation could be either in a site near theoperator-distal end of the galE gene, which signals rho factor termination oftranscription in vitro or in a site in the operator-promoter region. BacteriophageU3 was used to isolate galE mutations in HC2142 (a mutant exhibiting reducedresponse to capR control). P1 transduction was used to cross these mutants witha set of galE gene deletion. Analysis of the resulting Gal+ recombinantsindicates that the regulatory site is in the operator-promoter region. Hence, it isunlikely that capR functions in control as an anti-rho factor at the operator-distalend of the galE gene, but more likely as previously suggested, at a secondoperator distinct from one responding to galR repressor control. Upon inductionwith D-fucose, a promoter mutant (UV211) isolated previously expressed 20 to30% of the galactose enzymes that the wild type exhibited in the presence of theinducer D-fucose. The effects of various mutations in cya, capR, and galR ongalactokinase synthesis in this mutant were determined. Galactokinase wasderepressed by capR as well as galR, but the presence or absence of the cya geneproduct was unimportant.

Genetic and biochemical evidence establishedthat capR9 and capT mutations caused dere-pression of the galactose operon (11, 19, 20). Anew type of mutant with a lesion in the galactoseoperon has been isolated that exhibits reducedresponse to capR and capT gene control. Themutation was closely linked to galK (12).The capR9 (lon) mutation has a number of

pleiotropic effects on the cell. capR9 mutantsare sensitive to ultraviolet (UV) and X mysforming long filaments that eventually result indeath of the cell (3, 10, 21, 23, 39). The samemutation causes a mucoid phenotype and dere-pression of the enzymes involved in the biosyn-thesis of polysaccharide (5, 16, 17, 20, 22). BothUV sensitivity and the mucoid phenotype arethe result of a defect in a single cistron (7).CapR9 mutation also affects gene expression ofphage X and P1 (37, 40; Gayda, S.-S. Hua, Berg,and A. Markovitz, unpublished data). Morerecently Shineberg and Zipser have demon-strated that lon mutations are sufficient toenhance the stability of f3-galactosidase non-sense fragments (35). The deg mutants isolatedhave the same phenotype as capR9 mutants (6,35).

The gene product of capR has not beenidentified. Its biochemical interaction with thegalactose operon thus is not clear. DeCrom-brugghe et al. (8) demonstrated a rho-sensitivesite at the end of the galE gene using an in vitrotranscription system for the galactose operon.Examination of the biosynthetic pathway forpolysaccharide synthesis (Fig. 1) indicates thaturidine diphosphate (UDP)-galactose-4-epimer-ase (epimerase; EC 5.1.3.2), specified by thegalE gene in the gal operon, is one of theimportant enzymes for the biosynthesis ofUDP-galactose and UDP-glucose. Based on this func-tion, capR9 may specify anti-rho factor thatfunctions at or near the operator-distal end ofthe galE gene to cause partial derepression ofthe entire galactose operon. Alternatively, we(12) proposed that the new mutation in the galoperon, galO42, lies in the operator-promoterregion (Fig. 2). Deletion mapping allowed dis-crimination between the operator distal andoperator proximal sites.The new model for regulation of the gal

operon proposed that there are two promoters inthe gal operon, one cyclic adenosine 5'mono-phosphate (cAMP)-dependent and one cAMP-

510

on April 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

REGULATION OF GALACTOSE OPERON

GLUCOSE

G-6-P

/1y 171F-6-P

11]M-6-P

UDP-GALACTOSE

1911 OU[111UTP

G-I-P U-UDP-GLUCOSE CAPSULAR8 [101 2 CN POLYSACCHARIDE

UDP-GLUCURONICACID

GDP- L- FUCOSErarn

1(21 41 TPNH- (35 EPIM )rp GDP -0_D4KET0-MI- pTP MANNOSE D- RHAMNOSE

FIG. 1. Postulated biosynthetic pathway for capsu-lar polysaccharide synthesis in E. coli K-12. Thenumber in brackets refer to enzymes named asfollows: (1) phosphomannose isomerase; (2) phos-phomannomutase; (3) (GDP) guanosine 5'-diphos-phate mannose pyrophosphorylase; (4) GDP-mannosehydrolyase; (5) GDP-fucose synthetase; (6) phospho-glucose isomerase; (7) phosphoglucomutase; (8) UDP-glucose pyrophosphorylase; (9) UDP-galactose 4-epimerase; (10) UDP-glucose dehydrogenase; (11)polysaccharide polymerases.

independent (12). We are seeking more bio-chemical evidence to support such a model.Upon induction with D-fucose, a promoter mu-tant (UV211) isolated previously expressed 20 to30% of the galactose enzymes that the wild typeexhibited in the presence of the inducer D-fucose(2, 33). The effect of mutations in cya, capR,and galR on galactokinase synthesis in thispromoter mutant were determined.

Other possible mechanisms of capR regula.tion of galactokinase were studied.

MATERIALS AND METHODSBacterial and bacteriophage strains. The bacte-

ria utilized in this study are listed in Table 1.Bacteriophage U3 (41) and bacteriophage C21 (33)were provided by K. Paigen and S. L. Adhya andgrown as described. All strains of bacteria are deriva-tives of Escherichia coli K-12.

Media. M-9 minimal medium (1) was used to growcells for enzymatic assay. Either 0.6% glucose, 1%glycerol (vol/vol) or 0.6% galactose was used as carbonsource. These media were solidified by adding 1.5%agar. L-broth supplemented with 2 mM CaCl2 wasused for P1 transduction (11).

Chemicals. Adenosine triphosphate (ATP), D-fucose, and D-galactose were purchased from SigmaChemical Co. [4C ]galactose was purchased fromAmersham/Searle Nuclear-Chicago Corp. WhatmanDE81 cellulose paper was purchased from ReeveAngel Co. Omnifluor was obtained from New EnglandNuclear.

Genetic methods. Transduction was performed asdescribed by Lennox using bacteriophage Plkc (14). A0.1 ml volume of 25% sodium citrate was spread on theselection plates to prevent reabsorption of P1 phageand to increase the number of transductants.Enzyme assays. Cells were grown in minimal

medium in test tubes on a rotary drum or with

reciprocal shaking. Late-exponential-phase cells wereharvested for measurement of enzyme activity. Cellextracts were prepared by sonically disrupting thecells as described previously (11, 12). Galactokinasewas assayed according to the method of Sherman andAdler (34). Galactose and galactose-1-phosphate wereseparated by Whatman DE81 cellulose paper (42).Radioactive samples and protein (18) were measuredas described previously (12). Michaelis constants(Kin) were determined by Lineweaver-Burk reciprocalplots.

Selection of galE mutants with bacteriophageU3 for mapping of galO42. Bacteriophage U3 willonly absorb and lyse host cells that have galactose inthe cell wall (41). Cells that do not have galactose inthe cell wall are resistant to U3 phage. These U3-resistant cells have mutations in either galE, galU, orin a step to incorporate galactose into the cell wall(41). We were interested in obtaining galE mutationsin a galO42 genetic gackground. We chose the strainHC2142 (capR9, galO42) as the starting strain sincethe clones are mucoid. We have shown that some galEand galU mutations will render the cells nonmucoidon minimal glucose plates and these mutants will notgrow on minimal galactose plates (5, 11; Hua andMarkovitz, unpublished data). These observationspermitted us to identify tight galE or galU mutantsalthough a large number of U3-resistant cells areleaky. To distinguish galE from galU, we used F'8 galfor complementation tests. Fourteen galE clones wereisolated from 1,000 U3-resistant clones. The proper-ties of these mutants are summarized in Table 2.Among the 14 galE mutants, 10 are polar.

RESULTSIntragenic recombination of galO42, galE

mutants with galE deletion mutants. Themap position of galO42 is important in deter-mining the mechanism of capR regulation of thegal operon. To determine whether galO42 wasoperator proximal or not galO42 was mappedagainst deletions using the galE mutationsisolated in the strain containing galO42 (Mate-rials and Methods). Three deletion mutantswere chosen to map the set of galE, galO42

gol E gol T gal K

~~~ - -_

DNA Sites prc ocopR --p; QgolRfDNA

RNA copR+ c-AMP gol R+polymerose repressor CRP repressor

Interocting and/or copT +molecules activator? RNA

polymerose D-Fucose

FIG. 2. A modified model for the structure of thegalactose operon. Taken in part from Hua and Marko-vitz (12), capT products is proposed as a possibleactivator.

511VOL. 122, 1975

4

on April 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

HUA AND MARKOVITZ

TABLE 1. Bacterial and bacteriophage strains

Mutant allelesStrain important to Derivative, source, and/or genotype

this study

capR9

galO42 capR9

capR9 galU galK2

galO+galO+ capR9gal- A5061

gal- A503

gal- A105

gal- AS165

cyaA galP+galR-, cyaA, galP211galR-, galP211cyaA, galP211galP211galR-cyaA, galP211 capR9

galR-, cyaA, galP211, capR+galR-, galP211, capR9

galR-, galP211cyaA, galP211, capR9

cyaA, galP211galP211, capR9

galP211capR9galR-

galR-capR9

lysAgalRAlysAgalRA

lysAgalRAcapR9

F'gal+

cya+galP+

cyaAgalP+

F-, leu-6, proC34, purE38, tryE43, thi-1, aral4, lacYl,xyl-5, mtl-1, tonA23, tsx-67, azi-6, str-109, A-, galK2

P1 (proC+ capR9) x MC100; selected for Pro+; score formucoid phenotype (23)

P1 (ga1O42) x MC102; selected for Gal+ (12)P1 (gal+) x MC100; selected for Gal+P1 (trp+galU) x MC102; selected for Trp+; score fornonmucoid phenotype (galU) (ref. 11)

P1 (MC169) x MC129; selected for Pro+ nonmucoid (12)P1 (MC169) x MC129; selected for Pro+ mucoid (12)F-, his-, str-, XC1857, galUA5061; deletion of part of galE,

(33), from J. ShapiroF-, his-, str-, (XC1857), gal-A503; deletion of part ofgalE.

(33), from J. ShapiroF-, his-, str-, gal- AGA105; deletion of galK, galT and

part of galE. (33), from J. ShapiroF-, his-, str-, gal- AS165; deletion of galE, gaIT and part

of galK, (33), from J. ShapirolacI, glyR-, cyaA, from W. EpsteinF-, proC-, cyaA, galR-, galP211, from W. EpsteinF-, proC-, galR-, galP211, from W. EpsteinF-, proC-, cyaA, galP211, from W. EpsteinF-, proC-, galP211, from W. EpsteinP1 (MC169) x LU118, selected for Pro+, score capR9 byUV sensitivity and mucoid phenotype

P1 (MC1691 x LU118, selected for Pro+ nonmucoid.P1 (MC169) x LU119, selected for Pro+, score capR9 byUV sensitivity and mucoid phenotype

P1 (MC169) x LU119, selected for Pro+ nonmucoidP1 (MC169) x LU120 selected for Pro+, score capR9 byUV

sensitivity and mucoid phenotypeP1 (MC169) x LU120, selected for Pro+ nonmucoidP1 (MC169) x LU121 selected for Pro+ score capR9 by UV

sensitivity and mucoid phenotypeP1 (MC169) x LU121, selected for Pro+ nonmucoidP1 (gal+) x MC102; selected for Gal+ (11)P1 (thy+galR-) x MC129 thy; selected for Thy+, score

galR-(11)P1 (thy+galRj) x HC1002 thy-; selected for Thy+ scoregalR-(11)

M. Gottesman (28)P1 (PG8) x MC129 thy-, selected for Thy+ score Lys- Gal

constitutive (lysAgalRA), constructed by J. BushP1 (PG8) x HC1002 thy-; selected for Thy+, score Lys-

Gal constitutive (IysAgalRA); constructed by J. Bushmet-lys-, proC, ilv-, lac-, galA5061. From W. EpsteinP1 (LU112) x LU90; selected for ilv+, score cya+ (Lac+ and

Avir-sensitive phenotype) and then P1 (LU112), selectedfor Gal+

P1 (LU112) x LU90; selected for ilv+, score cyaA (Lac-and Avir-resistant phenotype) and then P1 (LU112),selected for Gal+

mutants. The extent of the deletions is shown in described in Table 3. Strain MS5061 contains aFig. 3 (33), as well as the mapping of the phage deletion of the entire operator-promoter regionU3-resistant galE mutants. All the phage U3- of the gal operon and extends the shortestresistant mutants in Fig. 3 were mapped as distance into galE of any available mutant.

MC100

MC102

HC2142MC129MC169

HC2001HC2101MS5061

MS503

GA105

S165

LU112LU118LU119LU120LU121HC2418

HC2318HC2419

HC2319HC2420

HC2320HC2421

HC2321HC1002HC1016

HC1017

PG8JBC1030

JBC1031

W4520/F' galLU90HC2301

HC2302

512 J. BACTERIOL.

on April 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

REGULATION OF GALACTOSE OPERON

TABLE 2. Characterization of galE negative mutantsa

Sensi- Galac-Stralin Frequency of tivity okinase Polarityreversion to ga- tins Pli

lactose,' activityeU5 3 x 10-7 r 0.03 YesU35

HUA AND MARKOVITZ

042 GENOTYPE PREDICTED

oll 042

both 042 and O'

042both 042 and 0

adl 042

b

FIG. 4a and b. Mapping of galO42 mutation withbacterial deletions. This is an analytical model ofrecombinational events which predicts the distribu-tion of galO+ and galO42 genotype among the Gal+transducts. MS5061, MS503, and GA105 are galEdeletion mutants which are described in Table 1.

that grew on revertants of strain U84. No Gal+clones were recovered when strain S165 was therecipient. Under the same conditions Gal+transductants were readily recovered whenstrains MS5061 and GA105 were the recipients.Such results suggest that reversion of strain U84was not an important consideration in theseexperiments. The following calculations lead tothe same conclusion. The reversion rate for thegalE mutants isolated was between 10- 6 to 10- 8.The mutants used for mapping had reversionrates from 10-v to 5 x 10-f (Tables 2 and 3).Usually 5 x 108 cells of the deletion mutantrecipient were spread on minimal galactoseplates. The regular gal+ or his+ transductionfrequency is approximately 10-1 per recipient.We calculate the Gal+ clones due to growth ofthe P1 on a revertant donor bacterium will be(10-7to5 x 10-f) x 10-6 x 5 x 108cells = 5 x10-5 to 2.5 x 10-4 gal+ transductants per his+transductant. The frequency of gal+ recombi-nants between galE, ga1042 and the set of galEdeletions was approximately 10-2 per his+transductant (Fig. 4) in the most importantrecombinations (U84 and U87). The interfer-ence due to reversion of galE mutants will rangefrom 1 to 10% of the recombinants.

Effect of cyaA, gaIR, and capR mutationson galactokinase synthesis by the gal pro-moter mutant, gaIP211. The galO42 mutationdefined a second operator region in the galoperon on the basis of its insensitivity to capR

regulation (12). The gal promoter mutation,galP211, reduces the basal level of gal operonenzymes (2, 33). We wanted to determinewhether or not galP211 might be of use indefining a second promoter if there were one.Isogenic strains with galP211 were provided byW. Epstein and constructed in this laboratory(Table 1). The effect of cyaA, galR, and capR ongene expression of galP211 were determined byassaying galactokinase level in these strains.The data are summarized in Table 4. The basallevel of galactokinase in the galP211 strain isapproximately 30% the level in the galP+ strain.galR causes a 10- to 15-fold derepression ofgalactokinase in galP211 which is similar to thederepression caused by galR in a galP+ strain(12). capR9 causes derepression of galP211 inboth a galR and galR+ strain. Derepression bygaIR or capR is not affected by the introductionof cyaA. In fact, cyaA does not affect enzymesynthesis in either galP211 or galP+ strains.Therefore, the site for gaiR and capR interac-tion has not been altered in galP211. The resultsin Table 4 do not permit one to decide whetherthere is one promoter or two promoters. Morepromoter mutants have to be isolated andcharacterized in order to provide a better under-standing to this problem.

Is gaiR repressor protein the target forcapR regulation? Previous results with a galRomutation in a capR+ and capR9 strain estab-lished that capR and galR regulate the galoperon independently. capR9 still causes two-

TABLE 4. Effect of various mutations on galactokinasesynthesis in gaIP211 promoter mutanta

Glucose GlycerolBacteria Pertinent asucon asStrain genotype" source carbon

source

HC2001 cya+, galP+, capR+ 0.23 0.21HC21PI1 cya+, galP+, cdpR9 1.14 1.02HC2321 cya+, galP, capR+ 0.060 0.047HC2421 cya+, gaiP, capR9 0.269 0.20HC2301 cya+, galP+ 0.21HC2302 cyaA, galP+ 0.25HC2302 cyaA, galP+ 0.23

(+cAMP)cHC2319 cya+, galP, galR-, capR+ 1.30 1.24HC2318 cyaA, galP, gaIR-, capR+ 1.73HC2419 cya+, galP, galR-, capR9 2.25 2.4HC2418 cyaA, galP, galR-, capR9 3.74HC2320 cyaA, galP, capR+ 0.072HC2420 cyaA, galP, capR9 0.356

a Cells were grown in M9 medium at 23 C. Enzyme activityis expressed as micromoles of galactose-l-phosphate formedper hour per milligram of protein at 37 C.

"The wild-type allele was present unless indicated other-wise.

cCells were grown in the presence of 2 x 10-a M cAMP.

gal E

RRU

04

042

----i--o

514 J. BACTERIOL.

on April 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

REGULATION OF GALACTOSE OPERON

fold derepression in a gaIR strain. However, allthe data do not eliminate the possibility thatthe galR repressor has two binding sites withinthe gal operator deoxyribonucleic acid (DNA).In the presence of the inducer, D-fucose, orgalactose, the conformation of the galR re-pressor would change so it would be releasedfrom one binding site. In capR9 cells, a smallmolecule could be produced that would alter theconformation of the galR repressor, so it wouldbe released from the other binding site. If thiswere the correct model for the action of capR9on gene expression of the gal operon, one wouldexpect that in a galRA mutant, capR9 wouldnot cause further derepression of the gal en-zymes. The results are summarized in Table 5.The derepression by capR9 is similar in the galRdeletion, and the gaIR point mutant. The par-ticular galRA comes from strain PG8. Thedeletion covers the entire lysA gene and part orthe entire gaIR gene. The exact extent of thedeletion of galR has not been determined (26,29). The results tentatively suggest that thecapR9 gene product does not interact with thegalR repressor protein directly.Other possible effects of capR9 on the

activity of galactokinase. In yeast, galactose-utilizing enzymes exist in aggregates; there is agene product which activates the enzyme com-plex upon exposure to inducer. Enzyme com-plex and activator associate with each other forthe catabolism of galactose (38). There is evi-dence in the phosphoenolpyruvate-dependentformation of D-fructose-1-P system for smallprotein factors that alter the Km and the Vmax ofthe enzyme (9). In E. coli and Salmonella,mutants in which one of the aminoacyl-transferribonucleic acid (RNA) synthetases is greatlydecreased in activity or increased in Km valuesare often derepressed in the synthesis of thecorresponding amino acid. This kind of regula-tion is common in eukaryotic cells (24). It hasbeen demonstrated that capR9 causes derepres-sion of gal messenger RNA synthesis (4). How-ever such a result does not entirely eliminatethe possibility that the capR9 mutation mayalter the apparent Km of galactokinase to itssubstrates and change the catalytic activity ofthe enzyme. The Km values for different sub-strates of galactokinase in capR+ and capR9cells were measured. The Km values for galac-tose were 6.6 x 10-1 M and 8.8 x 10-4 M incapR+ and capR9 cells. The Km values for ATPwere 2.1 0.4 x 104M and 1.6 4 0.6 x 10-4Mfor capR+ and capR9 cells. The differencesbetween Km values for galactose and ATP incapR+ and capR9 strains do not appear to besignificant.

TABLE 5. Effect of capR9 on the galactokinase level ina galRA mutanta

Galac-Bacterial strain Gene allele tokinase

(Amol/mg/h)

JBC1030 galRA, capR+ 1.6JBC1031 galRA, capR9 2.4HC1016 galR-, capR+ 1.5HC1017 galR-, capR9 2.5

a The cells were grown in M9 glucose at 23 C. Theenzyme was assayed at 25 C using a Dowex-50 columnto separate galactose from galactose-1-P. The protocolis as described (11).

DISCUSSIONThe results demonstrate that the galO42 mu-

tation, which exhibits reduced response tocapR9 (capT) control (12), lies in the operator-proximal portion of the gal operon and not atthe operator-distal segment of the galE gene.Hence, it is unlikely that capR (capT) functionsin control as an anti-rho factor at the operator-distal end of the galE gene, a possible model inview of the results of DeCrombrugghe et al. (8).The results support the model ofgalO42, being amutation in a second operator in the gal operon,as previously postulated (12). The capR geneproduct exhibits the properties of a repressorprotein in a number of biochemical genetic testsas well as UV radiation sensitivity tests (7, 13,20, 21, 23, 39). However, capT exhibits controlof enzymes of polysaccharide synthesis, includ-ing the gal operon, similar to capR (11, 12), andit is possible that the protein product of eithercapR or capT (or both) have an affinity for theDNA site in the gal operator defined by theoperator mutation galO42. The double mutantcapR9capT is no more derepressed than eitherof the single mutants (11). Although capR(lon)causes UV radiation sensitivity (10, 23), capThas no such effect (Markovitz, unpublisheddata). Thus capT only affects enzymes of poly-saccharide synthesis and may be the true regu-latory gene specifying a repressor or an activatorfor the gal operon and other operons of enzymesof capsular polysaccharide synthesis. If capTwere an activator then it would provide anexplanation for the following dilemma; strainswith a deletion of either the adenyl cyclase gene(cya) or the cAMP-binding protein (CRP) growon galactose and cAMP availability has littleeffect on gal enzyme levels (31). However invitro transcription (and translation) of the galoperon DNA is dependent on cAMP and CRP(27, 29, 42). The in vitro systems may bemissing the activator specified by capT. If in

VOL. 122, 1975 515

on April 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

HUA AND MARKOVITZ

fact capT gene product is an activator then wepredict the galO42 DNA should be transcribedin vitro in the absence of cAMP and CRP andextracts from a capT strain might containactivator and transcribe galO+ DNA in theabsence of cAMP and CRP. These experimentsare in progress.The frequency of recombinants was high in

the mapping of the short segment in the pro-moter region of the lac operon (25). Work byYanofsky et al. estimated that mutationalevents which are one nucleotide apart can bemapped at a frequency of about 0.02% (2 x10-i) in the tryptophan operon (44). If thedistance between two intragenic recombinationsites is 50 to 100 nucleotides, the frequency willbe 10-2 to 2 x 10- 2. The frequency we observedfor Gal+ recombinants between the galE, galO42mutant and galE deletion mutant is in thatrange. Recombination frequencies alone do notpermit ordering the galE mutants isolated withcertainty. Recombination values between dif-ferent mutants can be exceptionally high or low.It would appear that differences of single nu-cleotide pairs between donor DNA and recipientDNA can influence the frequency of recombina-tional events (36, 44). Therefore, the linearorder of sites presented (Fig. 3) is tentative.However, the mutants that gave no recombi-nants with the deletion recipients are moredefinitely placed as mapping under the particu-lar deletion (Fig. 3).There are some experiments that can be

interpreted using a two-promoter model for thegal operon.

(i) The binding of ribonucleate (RNA) po-lymerase to the gal promoter(s) indicates thatthere are two types of binding (43). In thepresence of CRP and cAMP six molecules ofRNA polymerase are bound to Xgal DNA invitro. In the absence of CRP and cAMP there isone molecule ofRNA polymerase bound to XgalDNA in the presence of ATP and guanosine5'-triphosphate and none if ATP and guanosine5'-triphosphate are not added (43). The datacan be interpreted as evidence for two RNApolymerase initiation sites, although one site atwhich affinity is altered by cAMP and CRP isan alternative interpretation.

(ii) Acridine dyes are known to intercalatebetween the bases in the DNA double helix,causing an extension and unwinding of thedeoxyribose backbone. There is experimentalevidence that an alteration in conformation ofthe DNA occurs at localized chromosomal re-gions in the presence of the dye (32). In the caseof the lac operon, which is a catabolite-sensitiveoperon, induction of lac nessenger RNA tran-

scription by the inducer isopropylthio-fl-D-galactoside probably involves a conformationalchange in the lac promoter region, which maythen exhibit increased affinity for acridineorange. This probably results in a decreasedaffinity for the formation of a transcriptioninitiation complex and decreased synthesis off-galactosidase. Excess cAMP partially re-versed the inhibition, presumably by displace-ment of the dye molecule and by favoring theformation of the transcription complex (32).Our preliminary results showed that inductionof the gal operon by D-fucose in the presence ofacridine orange caused a decrease in the synthe-sis of galactokinase. The inhibition was par-tially prevented by cAMP. However, cAMP didnot stimulate galactokinase synthesis in theabsence of acridine orange (Hua and Markovitz,manuscript in preparation); thus, treatmentwith acridine orange revealed a cAMP sensitivesite that may be in the gal promoter region (30).

Strain UV211 is the only gal promoter mutantisolated and characterized (33). Although galac-tokinase was derepressed by capR as well asgalR in strain UV211 the presence or absence ofthe cya gene was unimportant. The results dopermit one to decide whether there is onepromoter or two promoters.

ACKNOWLEDGMENTSWe thank James Shapiro for providing strains (Table 1)

and stimulating discussions. Wolfgang Epstein and JohnBush kindly constructed several strains as indicated in Table1. We would like to thank P. Bacha for competent technicalassistance.

This work was supported by Public Health Service grantAI 06966 from the National Institute of Allergy and InfectiousDiseases and American Cancer Society grant VC 116.

LITERATURE CITED1. Adams, M. H. 1959. Bacteriophages, p. 445-447. Intersc^-

ence Publishers, Inc., New York.2. Adhya, S. L. and J. A. Shapiro. 1969. The galactose

operon of E. coli K-12 I. Structural and pleiotropicmutations of the operon. Genetics 62:231-247.

3. Adler, H. I., and A. A. Hardigree. 1964. Analysis of a genecontrolling cell division and sensitivity to radiation inE. coli. J. Bacteriol. 87:720-726.

4. Buchanan, C. E., S.-S. Hua, H. Avni, and A. Markovitz.1973. Transcriptional control of the galactose operon bythe capR(lon) and capT genes. J. Bacteriol. 114:891-893.

5. Buchanan, C., and A. Markovitz. 1973. Derepression ofuridine diphosphate-glucose pyrophosphorylase (galU)in capR(lon), capS, and capT mutants and studies onthe galU repressor. J. Bacteriol. 115:1011-1020.

6. Bukhari, A. I., and D. Zipser. 1973. Mutants of Esche-richia coli with a defect in the degradation of nonsensefragments. Nature (London) New Biol. 243:238-241.

7. Bush, J. W., and A. Markovitz. 1973. The genetic basisfor mucoidy and radiation sensitivity in capR(lon)mutants of E. coli K-12. Genetics 74:215-225.

8. DeCrombrugghe, B., S. Adhya, M. Gottesman, and I.Pastan. 1973. Effect of rho on transcription of bacterial

516 J. BACTERIOL.

on April 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

VL127REGULATION OF GALACTOSE OPERON

operons. Nature (London) New Biol. 241:260-264.9. Hanson, T. E., and R L. Anderson. 1968. Phosphoenol-

pyruvate-dependent formation of D-fructose-1-phos-phatase by a four component phosphotransferase sys-tem. Proc. Natl. Acad. Sci. U.S.A. 61:269-276.

10. Howard-Flanders, P., E. Simson, and L. Theriot. 1964. Alocus that controls filament formation and sensitivityto radiation in Escherichia coli K-12. Genetics49:237-246.

11. Hua, S.-S., and A. Markovitz. 1972. Multiple regulatorygene control of the galactose operon in Escherichia coli.J. Bacteriol. 110:1089-1099.

12. Hua, S-S., and A. Markovitz. 1974. Multiple regulation ofthe galactose operon-genetic evidence for a distinctsite in the galactose operon that responds to capR generegulation in Escherichia K-12. Proc. Natl. Acad. Sci.U.S.A. 71:239-246.

13. Kang, S., and A. Markovitz. 1967. Induction of capsularpolysaccharide synthesis by p-fluorophenylalanine inE. coli wild type and strains with altered phenylalanylsoluble ribonucleic acid synthetase. J. Bacteriol.93:584-591.

14. Lennox, E. S. 1955. Transduction of linked characters ofthe host by bacteriophage P1. Virology 1:190-206.

15. Lieberman, M. M., C. E. Buchanan, and A. Markovitz.1970. Derepression of GDP a-D-mannose and UDP-glucose pyrophosphorylases by a regulator gene muta-tion; episomal dominance in partial diploids. Proc.Natl. Acad. Sci. U.S.A. 65:625-632.

16. Lieberman, M. M., and A. Markovitz. 1970. Derepressionof guanosine diphosphate-mannose pyrophosphorylaseby mutations in two different regulator genes involvedin capsular polysaccharide synthesis in EscherichiaK-12. J. Bacteriol. 101:965-972.

17. Lieberman, M. M., A. Shaparis, and A. Markovitz. 1970.Control of uridine diphosphate-glucose dehydrogenasesynthesis and uridine diphosphate-glucuronic acid ac-cumulation by a regulator gene mutation inEscherichia K-12. J. Bacteriol. 101:959-964.

18. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. F.Randall. 1951. Protein measurement with the Folinphenol reagent. J. Biol. Chem. 193:265-275.

19. Mackie, G., and D. B. Wilson. 1972. Regulation of the galoperon of Escherichia coli by the capR gene. J. Biol.Chem. 247:2973-2978.

20. Markovitz, A. 1964. Regulatory mechanisms for synthesisof capsular polysaccharide in mucoid mutants of Esch-erichia coli K-12. Proc. Natl. Acad. Sci. U.S.A.51:239-246.

21. Markovitz, A., and B. Baker. 1967. Suppression ofradiation sensitivity and capsular polysaccharide syn-thesis in Escherichia coli K-12 by ochre suppressors. J.Bacteriol. 94:338-395.

22. Markovitz, M., M. M. Lieberman, and N. Rosenbaum.1967. Derepression of Phosphomannose isomerase byregulator gene mutations involved in capsular polysac-charide synthesis in Escherichia coli K-12. J. Bacteriol.94:1497-1501.

23. Markovitz, A., and N. Rosenbaum. 1965. A regulator genethat is dominant on an episome and recessive on achromosome. Proc. Natl. Acad. Sci. U.S.A.64:1084-1091.

24. Metzenberg, R. L. 1972. Genetic regulatory systems inNeurospora. In H. L. Roman (ed.), Annual ReviewsInc., Palo Alto, Calif.

25. Miller, J., K. Ippen, J. Scaife, and J. Beckwith. 1968. The

promoter-operator region of the lac operon of E. coli. J.Mol. Biol. 38:413-420.

26. Nakanishi, S., S. Adhya, M. Gottesman, and I. Pastan.1973. Studies on the mechanism of action of the galrepressor. J. Biol. Chem. 248:5937-5942.

27. Nissley, S. P., W. B. Anderson, M. E. Gottesman, R. L.Perlman, and I. Pastan. 1971. In vitro transcription ofthe gal operon requires c-AMP and c-AMP receptorprotein. J. Biol. Chem. 246:4671-4678.

28. Parks, J. S., M. Gottesman, R. L. Perlman, and I.Pastan. 1971. Regulation of galactokinase synthesis bycyclic adenosine 3',5'-monophosphate in cell-free ex-tracts of E. coli. J. Biol. Chem. 246:2419-2429.

29. Parks, J. S., M. Gottesman, K. Shimada, R. A. Weisberg,R. L. Perlman, and I. Pastan. 1971. Isolation of galrepressor. Proc. Natl. Acad. Sci. U.S.A. 68:1891-1895.

30. Perlman, R., and I. Pastan. 1969. Pleiotropic deficiency ofcarbohydrate utilization in an adenyl cyclase deficientmutant of E. coli. Biochem. Biophys. Res. Commun.37:151-157.

31. Rothman-Denes, L. B., J. E. Hesse, and W. Epstein.1973. Role of cyclic adenosine-3',5'-monophosphate inthe in vivo expression of the galactose operon ofEscherichia coli. J. Bacteriol. 114:1040-1043.

32. Sankaran, L., and B. M. Pogell. 1973. Differential inhibi-tion of catabolite sensitive enzyme induction by inter-calating dyes. Nature (London) New Biol. 245:257-260.

33. Shapiro, J. A., and S. L. Adhya. 1969. The galactoseoperon of E. coli K-12. II. A deletion analysis of operonstructure and polarity. Genetics 62:249-264.

34. Sherman, J. R., and J. Adler. 1963. Galactokinase fromE. coli. J. Biol. Chem. 238:873-878.

35. Shineberg, J. B. and D. Zipser. 1973. The ton gene anddegradation of ,B-galactosidase nonsense fragments. J.Bacteriol. 116:1469-1471.

36. Stadler, D. R. 1973. The mechanism of intragenic recom-bination. In H. L. Roman (ed.), Annual review ofgenetics. Annual Reviews, Inc., Palo Alto, Calif.

37. Takano, T. 1971. Bacterial mutants defective in plasmidformation: requirement for the lon+ allele. Proc. Natl.Acad. Sci. U.S.A. 68:1469-1473.

38. Tsuyumus, S., and B. G. Adams. 1974. Dilution kineticstudies of yeast populations: in vivo aggregation ofgalactose utilizing enzymes and positive regulator mol-ecules. Genetics 77:491-505.

39. Uretz, R. B., and A. Markovitz. 1969. Dominance ofultraviolet radiation resistance in partial diploids ofEscherichia coli K-12. J. Bacteriol. 100:1118-1120.

40. Walker, J. R., C. L. Usery, and J. S. Allen. 1973.Bacterial cell division regulation: lysogenization ofconditional cell division lon+ mutants of Escherichiacoli by bacteri6phage lambda. J. Bacteriol.113:1326-1332.

41. Watson, G., and K. Paigen. 1971. Isolation and character-ization of an Escherichia coli bacteriophage requiringcell wall galactose U3 phage. J. Virol. 8:669-674.

42. Wetekam, W., R. Staack, and R. Ehring. 1971. DNA-dependent in vitro synthesis of the galactose operon.Mol. Gen. Genet. 112:14-27.

43. Willmund, R., and H. Kneser. 1973. Different binding ofRNA polymerase to individual promoters. Mol. Gen.Genet. 126:165-175.

44. Yanofsky, C., B. C. Carlton, J. R. Guest, D. R. Helinski,and U. Henning. 1964. On the colinearity of genestructure and protein structure. Proc. Natl. Acad. Sci.U.S.A. 51:266-271.

VOL. 122, 1975 517

on April 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/