Vol. 150, No. 2JOURNAL OF BACTERIOLOGY, May 1982, p. 572-5810021-9193/82/050572-10$02.00/0

Transcription in Bacteria at Different DNA ConcentrationsGORDON CHURCHWARD,t HANS BREMER,* AND RY YOUNGt

The University of Texas at Dallas, Richardson, Texas 75080

Received 19 June 1981/Accepted 9 December 1981

The effect of changing the DNA concentration on RNA synthesis, proteinsynthesis, and cell growth rate was studied in Escherichia coli B/r. The DNAconcentration was varied by changing the replication velocity or by changingreplication initiation in a thymine-requiring strain with a mutation in replicationcontrol. The results demonstrate that changes in DNA concentration (per mass)have no effect on the cell growth rate and the rates of synthesis (per mass) ofstable RNA (rRNA, tRNA), bulk mRNA, or protein or on the concentration ofRNA polymerase (total RNA polymerase per mass). Thus, transcription in E. coliis not limited by the concentration of DNA, but rather by the concentration offunctional RNA polymerase in the cytoplasm. Changing the DNA concentrationdoes, however, affect fully induced lac gene activity, here used as a model forconstitutive gene expression. The magnitude of the effect of DNA concentrationon lac gene activity depends on the distribution of replication forks over thechromosome, which is a function of the replication velocity. Analysis of thesedata reinforces the conclusion that transcription is limited by the concentration offunctional RNA polymerase in the cytoplasm.

Is the rate of transcription initiation in bacte-ria limited by the concentration of DNA or bythe concentration of functional RNA polymer-ase? Furthermore, is the rate-limiting step thebinding of RNA polymerase to the promoter orthe formation of an active initiation complex?These questions are important for a quantitativeunderstanding of bacterial gene expression, es-pecially for the class of genes that are constitu-tively expressed. The growth rate-dependentvariation in the expression of such genes hasbeen called "metabolic regulation" (35).

It has been argued that such regulation resultswholly, or in part, from a limitation by the DNAconcentration (9, 16, 27) or, alternatively, byfunctional RNA polymerase (3, 6, 32). The ques-tion has eluded experimental solution because ofthe difficulty in manipulating the two parametersin vivo without changing additional parametersthat affect transcription (see below).We have previously studied the properties of a

mutant of Escherichia coli B/r, strain TJK16,which has a lower DNA concentration than itswild-type parent due to a mutationally alteredcontrol of initiation of chromosome replication(10, 12). TJK16 is auxotrophic for thymine, andits DNA concentration can be further loweredby growing it at low thymine concentrations(12). Therefore, this strain is useful for studying

t Present address: Universitd de GenFve, Departement deBiologie Moleculaire, CH-1211 GenEve 4, Switzerland.

t Present address: Texas A&M University, Department ofMedical Biochemistry, College Station, TX 77843.

the effects of varying DNA concentration on therate of in vivo transcription. The results suggestthat in both B/r and TJK16, DNA is transcribedunder conditions ofDNA excess; i.e., transcrip-tion is limited by the concentration of activeRNA polymerase rather than by DNA. In E. colionly 20 to 30% of the total RNA polymeraseenzyme is actively engaged in transcription(RNA chain elongation) at any given time,whereas 70 to 80% of the RNA polymeraseappears to be inactive (39). Since the resultshere indicate that DNA is not limiting the rate oftranscription, the inactivity of RNA polymerasemust be caused by other factors rather than alow DNA concentration.

MATERLALS AND METHODSBacterial strains and growth conditions. The bacteria

used were E. coli B/r A (ATCC 12407; 22) and E. coliTJK16, which is a thyA deoB derivative of E. coli B/rA obtained from J. Kwoh (M.S. thesis, University ofTexas at Dallas, 1975). NF955 ilv thr leu thi (A b515b519 c1857 S7 A dilv5), used to prepare X dilv DNA forhybridization studies, was kindly provided by N. Fiil.Growth conditions were as described previously (23,39).

Determination of cell mass, DNA, protein, and RNA.Cell mass (units of optical density at 460 nm [OD4w]per 109 cells) and DNA (colorimetric assay) weredetermined as described previously (7, 12). There wasno significant difference between the plating titer andthe particle concentration for either strain used. Todetermine protein and RNA, 5-ml samples were pre-cipitated with 1 ml of 3 M trichloroacetic acid, filteredthrough glass fiber filters (Reeve Angel; 984H),

572

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

GENOME CONCENTRATION AND GENE ACTIVITY 573

washed with tap water, dried, and placed in glass vials.After hydrolysis with 2.0 ml of 0.2 N NaOH at 23°C for20 h, 0.5 ml of the hydrolysate was removed forestimation of the protein content by a modification ofthe method of Lowry et al. (7, 28). The protein assays

were incubated for 20 h at 23°C. Under these condi-tions an absorbancy at 750 nm/OD40 ratio of 1 corre-sponds to 9.1 x 1017 amino acids per OD460 unit ofculture (calibration with bovine serum albumin). ForRNA, the remaining alkaline hydrolysate (after remov-al of the sample for estimation of protein) was treatedwith an equal volume of ice-cold 0.5 M perchloric acidto precipitate protein and DNA and filtered through aBact-T-Flex membrane filter (Schleicher & SchuellCo.; 0.45-p.m pore size). The absorbancy at 260 nm ofthe ifitrate was determined. An absorbancy at 260 nm/OD4w ratio of 1 corresponds to 4.5 x 1016 RNAnucleotides per OD4w unit of culture (calculated fromthe molar extinction of E. coli RNA at acid pH). Thereproducibility of the colorimetric assays for DNA,RNA, and protein is shown in Table 1.

Determination of 3-galactosidase activity. Enzymeactivity in fully induced cultures was determined byusing standard methods (15), except the reaction was

carried out at 23°C for 30 min.Determination of the fractional rate of stable RNA

synthesis. Samples (0.5 ml) of culture were removedand labeled with 5 ,ul of [5-3H]uridine (Schwarz/Mann;2.5 .Ci/5 ,ul, 21 Ci/mmol) for 1 min. The pulse wasterminated by the addition of lysing medium (5), andthe sample was immediately placed in a boiling waterbath for 1 min. The fraction of label in rRNA wasdetermined by hybridization to A ilv5, which carries anrRNA gene (13). Phage DNA was prepared afterinduction of NF955 by the method of Miller (31) andloaded onto Bact-T-Flex filters (0.45-p.m pore size, 10to 25 p.g/fflter) after alkaline denaturation. Hybridiza-tion was carried out in scintillation vials containing 1ml of 2x SSC (lx SSC is 0.15 M NaCl-0.015 Msodium citrate), 0.5% sodium dodecyl sulfate, 0.02 MEDTA, at 67C'for 18 h. Each vial contained two DNAifiters, a blank ifiter without DNA or with heterolo-

gous A or T5 DNA, approximately 0.05 ,ug of pulse-labeled RNA, and approximately 0.005 ,ug of purified(18) "C-labeled rRNA as a standard to determine thehybridization efficiency. In control experiments it wasfound that nonspecific hybridization to filters loadedwith A or T5 DNA was indistinguishable from bindingto blank filters. After corrections for background andhybridization efficiency the fraction of pulse label inrRNA (1,4to) was determined. The fractional rate ofsynthesis of stable RNA, rjrtot, was obtained from 41

l4,, by using the formula: r,/rt., = ( + [0.86- (4ltw)] x 0.86}. This formula takes into account thedifferent mole fractions of pyrimidines in stable andmRNA (0.43 and 0.5, respectively; ratio, 0.86) and thefraction of stable RNA that is rRNA (0.86). Thesynthesis rate of stable RNA, r,, includes the spacermaterial in the stable RNA precursor. The relative rateofmRNA synthesis, rSrtot, is equal to the difference 1- rjrtt,; r,0t is the total instantaneous rate of RNAsynthesis, including mRNA and unstable spacers instable RNA precursors.

Determination of RNA polymerase. Determination ofRNA polymerase was as described previously (39)from the amount of , and 1B' RNA polymerase subunitprotein found after sodium dodecyl sulfate-polyacryl-amide gel electrophoresis.

RESULTS

Two ways of changing the DNA concentration.The strain TJK16 is a derivative of E. coli B/rwhich, in addition to the thyA deoB markersconferring a low-thymine requirement, has athird mutation which increases its "initiationmass" (10, 12). Since TJK16 has a greater cellmass per replication origin, it has less DNA permass, i.e., a lower DNA concentration, in com-parison with the B/r parent at any growth rate(Fig. la). At high thymine concentration (>10p,g/ml), the DNA chain elongation in this strainis indistinguishable from that in the parent strain

TABLE 1. Reproducibility of colimetric assays for DNA, RNA, and proteinDNA RNA Protein

Dayawb Deviaiionbtion iion b DeviationDa~A60 from avg (%) 260 from avg (%) 750 from avg (%)

1 0.337 -2.0 1.244 -2.0 0.486 -8.00.354 +3.0 1.252 -1.3 0.507 -4.10.355 +3.3 1.258 -0.9 0.517 -2.20.360 +4.7 1.277 +0.6 0.510 -3.5

4 0.333 -3.1 1.221 -3.8 0.543 +2.70.340 -1.1 1.270 +0.1 0.574 +8.6

5 0.335 -2.6 1.284 +1.2 0.514 -2.70.355 +3.3 1.310 +3.2 0.527 -0.3

6 0.331 -3.7 1.260 -0.7 0.553 +4.60.338 -1.7 1.314 +3.6 0.554 +4.8

Avg or crc 0.344 a = 3.2% 1.269 a = 2.3% 0.529 a = 5.7%a Ten samples (10 ml each for DNA and 5 ml each for RNA and protein) were taken simultaneously from the

same culture, chilled on ice (B/r in glucose minimal medium, r = 43 min, OD460 = 0.98), precipitated withtrichloroacetic acid, and collected on glass fiber filters immediately, but further processed on different days asindicated.

b Blank values subtracted.I Standard deviation, a (%) = NMT/£(n - 1); A, deviation from average in percent; n, number of samples (10).

VOL. 150, 1982

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

574 CHURCHWARD, BREMER, AND YOUNG

- 0E,t I,ot X 120o0 E

V. . .

~~~1.0~~~~~~~ 0 C~20"

0 E 8 t Y S 3 n 'E1.0 ~A 0

U0 B/ 6 E

.2 0

b lo

0 160E..

~8 0 'N ~~~~~~~~~~~~E*0 Li~~~~~g ~6-o 0

ZO 0 60

36

TJK1groing n20,ugofthyin pe ml(-1r1,go8hmieprmO.Ec ontrpeet h vrg

CO C

0~~~~~~~~0

N 1 10C 06-

4 1

-0 1.0 2.0 3.0

FIG. 1. Concentration ofDNA and RNA in E. coli B/r and TJK16. Comparison of the concentrations ofDNA(DNA/OD4w unit, panel a), RNA (RNA per OD4w unit, panel b; RNA per protein, panel c) in E. coli B/r (0) andTJK16 growing in 20 ~L of thymine per ml (A) or I p.g of thymine per ml (EJ). Each point represents the averageof four values from one culture (two duplicate samples taken a doubling time apart). For estimation of theabsolute DNA concentration it was assumed that 1 OD4w unit corresponds to 0.3 ,ul of cell volume (12). Forconversion of the RNA/protein ratio into the fractional rate of synthesis of ribosomal protein, ar (24), it wasassumed that 83% of the bacterial RNA is rRNA (14% tRNA [17, 41] and 3% mRNA [26, 33]), and that a 70Sribosome contains 6,300 amino acid residues in protein and 4,720 nucleotide residues in rRNA.

(11), i.e., the thyA deoB mutations have noeffect. This implies (14) that at high thyminelevels the distribution of replication forks on thechromosome and the relative gene dosages(number of gene copies per genome equivalentof DNA) are the same in B/r and TJK16. Onlythe effects of the initiation mutation are thenapparent under these conditions, i.e., the con-centration of all genes in TJK16 is reduced bythe same factor.

In a "step-up" experiment the velocity ofreplication forks is increased by increasing theconcentration of thymine in the medium from aninitial low concentration (34). This changes thedistribution of replication forks over the chro-mosome, i.e., the chromosomes become lessbranched because rounds of replication are com-pleted faster. As a result, genes far from thereplication origin increase in concentration, butgenes near the origin do not (2, 9). Figure lashows that average gene concentrations (DNA/

OD40 unit) in TJK16 increased with increasingthymine concentration.The concentration ofDNA in bacteria is also a

function of growth rate (9, 21); in E. coli B/r andin the Thy- initiation mutant TJK16, the DNAconcentration decreases with growth rate (ob-served above 0.6 doublings per h) and reaches aminimum level above 1.5 doublings per h (Fig.la). Assuming that 1 OD40 unit of cells has avolume ofabout 0.3 pl (12), intracellular concen-trations ofDNA are estimated to vary between 5and 25 mg/ml depending on the growth rate andreplication velocity (Fig. la).RNA polymerase concentration and activity.

The amount of RNA polymerase 1B and I' sub-unit protein was determined for E. coli B/r andTJK16 growing in glucose-amino acids mediumat high and low thymine concentration (Table 2).Assuming that each subunit was part of a com-plete polymerase core enzyme, the fractionalrate of accumulation of RNA polymerase, aop,,

J. BACTERIOL.

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

GENOME CONCENTRATION AND GENE ACTIVITY 575

TABLE 2. Accumulation of RNA polymerase (ap =RNA polymerase protein per total protein) in E. coli

B/r and TJK16 growing in glucose-amino acidsmedium and, in the case of TJK16, supplementedwith 20 or 1 ,ug of thymine per ml as indicated

Thymine Doubling ProteinStrain concn time concna opb (%) Avg ap

(pLg/ml) (min) (mg/ml)

B/r 0 25 2.36 1.2827 1.59 1.29 1.2927 1.89 1.30

TJK16 20 30 2.22 1.3930 2.08 1.16 1.3528 2.13 1.49

TJK16 1 30 2.74 1.2531 2.44 1.31 1.3428 1.94 1.46

1[l33]c]B/r 0 25 2.36 ] 1 2 1.281.l22]

a Protein concentration in each sample used forelectrophoresis was determined by the method ofLowry calibrated with bovine serum albumin.

b a1p, was calculated as micrograms of core RNApolymerase x 100 per microgram of total protein;average of six values (two slots on three gels) for eachculture (one horizontal line). Each culture growth wasrepeated three times resulting in slightly differentdoubling times. Each gel also contained two slots witha known amount of bovine serum albumin for calibra-tion.

c The reproducibility of the estimate from gel to gelis shown by the six values which resulted in theaverage ap value of 1.28%. The brackets indicate thetwo slots from one gel.

was found to be about 1.33% (range, 1.29 to1.35% [percentage of total protein that is RNApolymerase]) for both strains and independent ofthymine concentration. Since mRNA and stableRNA synthesis per OD460 unit were the same forboth strains (see below), it seems likely that alsothe concentrations of actively transcribing RNApolymerase molecules are the same.The observation that RNA polymerase syn-

thesis was independent ofDNA concentration isimportant for models of RNA polymerase gene(rpoBC) control; it means that either the concen-tration of the presumed control factor (37) is ingreat excess over its binding sites on the DNA(see below), or the activity of this factor itself iscontrolled.RNA concentrmtion and stable RNA synthesis.

The concentration of RNA (amount of totalRNA per OD4w unit) increased with growth ratein both strains; at a given growth rate, the RNAconcentrations were not significantly different inB/r and TJK16 (Fig. lb). Most (-97%) of thebacterial RNA is stable (26, 33); under condi-

tions of exponential growth and for doublingtimes of less than 60 min, this stable RNA is 86%rRNA and 14% tRNA (17, 41). Thus, at a givengrowth rate, the rates of rRNA and tRNA syn-thesis per OD40 unit were independent of DNAconcentration. However, the synthesis rates pergene did depend on DNA concentration: inTJK16, with all genes (including rRNA genes) atlower concentrations as compared with B/r, thetranscription per gene was correspondinglyhigher (Fig. 2). The transcription of rRNA geneswas also a strong function of growth rate, in-creasing from 3 to 30 (B/r) or 12 to 50 (TJK16)transcriptions per rRNA gene per min over atwo- to threefold range in growth rate (Fig. 2).Ribosome synthesis and function. The amount

of RNA per protein (ratio R/P, proportional tothe number of ribosomes per amount of protein)is an indicator of ribosome synthesis and func-tion (Fig. ic; ar = ribosomal/total protein);obviously the denominator of the ratio RIP rep-resents the product of ribosome function (ribo-somes make protein), whereas the numerator isessentially a measure of the number of ribo-somes. RNA per protein was found to increasewith growth rate in an identical manner in wild-type E. coli B/r (Fig. lc) and in TJK16, both athigh and at low levels of exogenous thymine(Fig. lc). This means that ribosome synthesisand function are independent of DNA replica-tion, at least within the limits represented bywild-type B/r and the replication mutant TJK16.mRNA synthesis. The relative proportions of

stable and mRNA synthesis were the same in B/rand TJK16 (Table 3); thus, bulk mRNA synthe-

04, 50

40 -

C30 TJKI16

.2

20 -

0 B/rC) 10

0 10 20Growth rate (doubl./h)

FIG. 2. Transcription of rRNA genes in B/r andTJK16 (at any thymine concentration), calculatedfrom the rate of rRNA synthesis in nucleotides perminute per genome [(drRNA/dt)/genome = RNA/genome x 0.83(1n2/T); RNAlgenome is the ratio ofcurves in Fig. lb and a; 0.83 is the fraction of totalRNA that is rRNA, assuming 14% tRNA (17, 41) and3% mRNA (26, 33); r is the doubling time] and fromthe number and map locations of the seven rRNAgenes (19), using the formula for the relative genedosage from reference 2 and from the number ofnucleotides per ribosome (4,720).

VOL. 150, 1982

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

576 CHURCHWARD, BREMER, AND YOUNG

TABLE 3. Relative proportions of stable and mRNA synthesis in E. coli B/r and TJK16 growing in glucoseminimal medium (TJK16 supplemented with 20 or 1 Rg of thymine per ml, as indicated)

Total cpm in hy- Relative proportions of:bridization mix

Thymine Doubling HybridizationStrain concn time efficiencyc Stable RNA mRNA

(Ijg/ml) (min) 3Ha 14cb (%) synthesis synthesis%d Avg % (avg %)

B/r 0 48 5616 258 64 52

46 { 6029 258 34 51 4 466029 258 50 55145 6573 290 22 57

TJK16 20 45 5735 380 82 49 48 52e46 5963 368 84 46 4

1 45 2638 380 85 55 546 3125 368 92 481 52 48

a 3H-pulse-labeled RNA.b ['4C]rRNA (purified) added to hybridization mixture for determination of the hybridization efficiency.c The differences in hybridization efficiency are mainly due to differences in the amount of A dilv DNA loaded

to the filters. In the B/r experiments, about 10 Fg ofDNA was used per filter; in the TJK16 experiments, about 25p.g of DNA was used.d Each value given is the average from three hybridization vials.e The slightly higher mRNA values for TJK16 in comparison with B/r mean that the rate of total RNA

synthesis in TJK16 is 2 to 6% higher than in B/r (despite the lowerDNA concentration in TJK16). This differenceis experimentally not significant.

sis, like stable RNA synthesis per OD4w unit,was not affected by changes in DNA concentra-tion.To focus on a particular mRNA, we compared

lac transcription in the two strains. Most mRNAsynthesis is specifically regulated; to observeonly the effects of gene concentration and poly-merase availability, we used conditions wherethe lac operon was physiologically constitutive(i.e., maximally induced). Since the position ofthe lac genes is midway in the replication path(1), there is always about 1 lac gene per genomeequivalent, regardless of growth rate. (This isnot true for genes near the origin or the terminusof replication.) Thus, measuring the relativeamount of DNA per OD4w unit is the same asmeasuring the relative concentration of lacgenes. The synthesis of lac mRNA was deter-mined indirectly from the synthesis of totalmRNA and from the synthesis of P-galacto-sidase enzyme (see below). The results depend-ed on the kind of experimental condition thatwas used to change theDNA concentration, i.e.,whether it was due to a change in replicationvelocity or due to a change in replication initia-tion. These two conditions are examined sepa-rately below.

(i) Gene activities during a step-up. A replica-tion velocity step-up was found to have no effecton the rate of synthesis (per mass) of stableRNA (Fig. 3b), bulk mRNA (Fig. 4b), andprotein (Fig. 3b). Constitutive ,B-galactosidasesynthesis per mass increased by 40%, comparedwith the control culture (Fig. 4a), in agreementwith previous observations (9). Since the syn-

thesis per OD40 unit of 3-galactosidase is essen-tially the fractional rate of lacZ transcription pertotal mRNA transcription, the apparent increasein lac expression per OD460 unit actuallystemmed from a constant rate of transcriptionper lac gene and an increased number of lacgenes per mass (see below) (Table 4).

Since the genes for rRNA are clustered nearthe origin of chromosome replication (1), andsince a step-up does not affect replication of theorigin (4, 34), the concentration of rRNA genesand the rate of rRNA synthesis per rRNA generemain unaltered by the step-up (Fig. 2; thecurve for TJK16 is valid for any thymine con-centration). Thus, changes in DNA concentra-tion brought about by changes in replicationvelocity affect neither rRNA nor constitutive lacgene activities (Table 4).

(Bi) Gene activities after a change in replicationinitiation. In Fig. 5, the ratios of lac gene con-centrations in B/r and TJK16 are shown for highthymine levels when the only difference betweenthe two strains was in the control of replicationinitiation. The mutant had from 50 to 70% of thelac gene concentration of the wild type, depend-ing on the growth rate. The P-galactosidasesynthesis per OD4w unit was nearly equal in thetwo strains, although the synthesis in the mutantrelative to the parent B/r decreased somewhatwith increasing growth rate. By dividing theenzyme synthesis per OD4w unit by the numberof lac genes per OD4w unit (-DNA/OD4w), oneobtains a measure for the relative transcriptionra,te per lac gene (see below). At all growthrates, constitutive transcription per lac gene was

J. BACTERIOL.

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

GENOME CONCENTRATION AND GENE ACTIVITY

02

En0E= 1.

a),

30 60 -60 -30 0Time after step-up (min)

E60Nu

C

40_4Z;z

cr

E.20 N-

00

P._0

10 °a-

X8

_

0r_

Ia

FIG. 3. Transcription in TJK16 (growing in glucose-amino acids medium + 1 mM IPTG) during a step-up inthe replication velocity brought about by an increase in thymine concentration from 1 to 20 ,ug/ml. (a) Massaccumulation; dilutions as indicated (dilution factors). (b) Accumulation of DNA, RNA and protein; DNAaccumulation (---) calculated (4).

30% higher in the mutant. A different extent ofcatabolite repression in the two strains was notinvolved; the ratio of the rates (not the absoluterate) was independent of the presence of 0.01 Mexogenous cyclic AMP (data not shown), shownin previous work (15) to be in excess of theamount necessary to maximally reduce catabo-lite repression. Thus, lac gene transcription in-creases when the concentration of all genes isreduced proportionately. Clearly, DNA concen-

0.

,o

0.4e 1.2=

1.0CLl$ 0.80

a 0.1

- a <

I I I I c

l l0 --40 0 +40

tration does not limit lac gene expression inbalanced growth.

DISCUSSIONEffects of DNA concentration on the total rate

of RNA synthesis. In the past, the DNA concen-tration in bacteria was found to vary with thegrowth rate or with the replication velocity (9,21), but in these cases the accompanying generegulation and changes in replication fork pat-

aO--

S

zgr 0.!4p.00C-c) ,

|- b A 0 r

n A I I I I I I I C)

-40 0 +40Time after step-up (min)

FIG. 4. lac gene expression and mRNA synthesis after a stepwise increase in the replication velocity (sameexperiment as in Fig. 3). (a) Change in the fractional synthesis of P-galactosidase (alac), compared with changesin mRNA synthesis per genome (---) (calculated from Fig. 3b and 4b) and in the lac gene concentration (lac genes/OD4w unit) (- * *). (The number of lac genes per OD4w unit is equivalent to genome equivalents per OD4 unitfrom Fig. 3a and b.) (b) Changes in the relative synthesis rates of stable and mRNA, r,/rtot and rm/rtot.

VOL. 150, 1982 577

"ID

* .V

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

578 CHURCHWARD, BREMER, AND YOUNG

TABLE 4. Effect of changing DNA concentration on macromolecular synthesis rates and cell composition inE. coli B/r growing in glucose-amino acids medium at 2 doublings per h

n-fold change in:

DNA concnConcn Activity Rate per mass

changed by: MNarNAdobln RNA/ Protein/DNAe pol-RNA rRNA lac rRNA lac Bulk rRNA, Pro- doubli genomea genomeamDeayseb genec gened gene' genef ImRNA9 tRNAa tein' time

Initiation 0.6 1.0 0.6 0.6 1.7 1.3 1.0 1.0 1.0 1.0 1.7 1.7control

Replication 0.7 1.0 1.0 0.7 1.0 1.0 1.0 1.0 1.0 1.0 1.5 1.5velocitya Data from Fig. 1.b Data from Table 2.c rRNA gene concentration changes as much as DNA concentration (0.6-fold) by initiation control, and does

not change (1.0-fold) by changes in replication velocity (see text).d lac gene concentration changes as much as DNA concentration (see text).e rRNA gene activity changes as the quotient of (change in rRNA synthesis per mass)/(change in rRNA gene

concentration).f Data from Fig. 3, 4, and 5.8 Data from Table 3.

terns have obscured the effect of DNA concen-tration on transcription. In temperature-sensitive DNA initiation mutants, the DNA con-centration can be altered by growing the bacteriaat semipermissive (i.e., intermediate betweenfully permissive and nonpermissive) tempera-tures (20) without changes in replication forkpatterns, but the change in temperature affectsthe rate of transcription, again making the re-sults difficult to interpret. In the current work,we have utilized a strain with a non-tempera-ture-sensitive DNA initiation mutation in which

m-

Go

-

0

0

1.

0 1.0Growth rote (doubl./h)

FIG. 5. Comparison of lac gene expression in E.coli B/r and TJK16 growing at 20 ,ug of thymine per ml.(-), lac gene concentration; the number of lac genesper OD4w unit was assumed to be equal to the numberof genome equivalents of DNA per OD4w unit (seetext) and was obtained from the DNA measurementsof Fig. 1. (0), ,B-galactosidase per OD4w unit, eachpoint represents the ratio of the P-galactosidase con-centrations obtained from two identical cultures, B/rand TJK16, maximally induced for 3-galactosidase.(---), lac mRNA per gene, the ratio lac gene activity inTJK16/lac gene activity in B/r corresponds to the ratioof the lower two curves (see text).

the DNA concentration is reduced in compari-son to its wild-type parental strain without con-comitant changes in growth rate or in replicationfork patterns (12). This mutant is thus ideallysuited to study the effects of altered DNA con-centration on the rate of transcription. The datain Table 4 demonstrate that total RNA synthesisin E. coli per unit mass is not affected by areduction in DNA concentration; we concludethat transcription in E. coli is not limited by theconcentration ofDNA (i.e., by the availability offree promoters). Under such conditions, freefunctional RNA polymerase would be increasedby a decrease in total DNA. The concentrationof free RNA polymerase can be monitored bythe transcription of constitutive genes; theexpression of maximally induced lacZ genesobserved here at different DNA concentrationsqualitatively agreed with this expectation andthus reinforced the conclusion that transcriptionin E. coli is limited by the concentration ofactive RNA polymerase rather than by DNA.

This conclusion apparently contradicts theresults obtained by Seeburg et al., using an invitro transcription system with purified E. coliRNA polymerase and defined DNA fragments ofphage fd containing only one promoter, theseworkers observed that 20 s to 3 min (half time)was required for formation of an active initiationcomplex after the initial RNA polymerase bind-ing (38). Using these data, Maal9Se estimated thata concentration of two to three free RNA poly-merase molecules per cell would suffice to satu-rate the bacterial promoters with polymerase; athigher enzyme concentrations, the rate of for-mation of the active RNA chain initiation com-plex would be rate limiting (30). Since freefunctional RNA polymerase in bacteria has not

I

lac mRNA/gene .

,0 - cggl/A460-8loc g e A6

o. I' '- I '-

J. BACTERIOL.

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

GENOME CONCENTRATION AND GENE ACTIVITY 579

been directly measured, the estimated value ofenzyme concentration required for promotersaturation does not help to answer the questionof whether bacterial transcription is limited byDNA or by RNA polymerase. The observation(Fig. 2) that rRNA promoters showed no signs ofsaturation and in fact underwent initiation up toonce per second, indicates that in vivo theformation of active initiation complex is at least1 order of magnitude faster than under the invitro conditions studied by Seeburg et al. (38),which supports our conclusion that DNA con-centration is not limiting the rate of transcriptionin E. coli.

It is important to note the units used in theseanalyses. If RNA synthesis were measured withother reference units rather than per unit mass,e.g., as rate per cell or rate per genome (29),changes in DNA concentration would affectRNA synthesis indirectly. This reflects the ef-fects of DNA replication on these referenceunits, i.e., on DNA and cell number, rather thaneffects on RNA synthesis. Such units must beavoided in a consideration of general transcrip-tional regulation.RNA polymerase activity. In the condition of

DNA excess, free polymerase should be a smallfraction of total enzyme, and the major portionof the enzyme should be engaged in transcrip-tion. By quantitating RNA polymerase subunitsin DNA-free minicells, Shepherd (Ph.D. thesis,The University of Texas at Dallas, 1980) foundthat the concentration of cytoplasmic RNApolymerase was indeed a small fraction of thetotal enzyme. Yet only 20 to 30% of all RNApolymerase in E. coli B/r is active in transcrip-tion at any given time (39). If RNA synthesis isnot limited by DNA, the inactivity of a largeproportion of bacterial RNA polymerase musthave other reasons than a low DNA concentra-tion. It is conceivable that the RNA polymeraseactivity is limited by the availability of sigmafactor. The molar amount of sigma per cell isabout 30% of that of core enzyme (24, 25).However, since sigma factor is only required forRNA chain initiation and is released from thetranscription complex immediately after initia-tion (8), it is not obvious why a sigma/core ratioof 0.3 should not be sufficient to activate allRNA polymerase, i.e., the ratio of free sigma tofree core enzyme is presumably much greaterthan 1. Another indication that factors otherthan DNA concentration may limit the RNApolymerase activity is the observation that theRNA polymerase activity rapidly responds tochanges in the intracellular level of guanosinetetraphosphate (36).Kawakami et al. (25) observed that up to 100%

of the total E. coli RNA polymerase could berecovered as nucleoid-bound enzyme when the

nucleoids were prepared from exponential bac-teria at a KCl concentration of 0.2 M; thisfraction decreased during the transitional periodbetween exponential growth and the stationaryphase. The fact that enzyme was DNA bound invitro does not necessarily mean that it is alsobound in vivo, where ionic conditions are differ-ent. Even if most polymerase enzyme wereDNA bound in vivo, which would be consistentwith the minicell data, this would not mean thatthe enzyme must be actively engaged in tran-scription. Thus, the data of Kawakami et al. arenot inconsistent with a low polymerase activityof 20 to 30%.

Expression of individual genes after changes inDNA concentration. If the expression of specificgenes is considered rather than total RNA syn-thesis, changes in DNA concentration can beexpected to produce more subtle effects since itthen becomes important to consider the locationof the gene on the chromosome with respect tothe origin of replication and the particular man-ner by which the DNA concentration waschanged; changes in DNA replication velocityalter relative gene dosages, depending on thegene location, whereas changes in replicationinitiation have no such effect. Further, massaction, just as it predicts changes in free RNApolymerase as a result of changes in DNAconcentration, also predicts changes in the con-centration of free positive or negative controlfactors that bind to DNA. These effects, as theyare relevant for the interpretation of the dataobtained here, will be discussed below.The strain TJK16 has all genes in a lower

concentration than its wild-type B/r parent, andsince the rate of total transcription per OD40unit is the same in TJK16 as in B/r (Table 4), thegenes producing the bulk of stable RNA andmRNA must be more active in TJK16 than in B/r. This is interpreted as a reflection of theincrease in the concentration of free (cytoplas-mic) RNA polymerase (3) due to the decreasedDNA concentration. The fact that the relativeproportions ofrRNA and mRNA synthesis werenot affected by changes in DNA concentration(Table 3) suggests that any differences in individ-ual gene activities do not reflect a guanosinetetraphosphate-dependent control ofRNA poly-merase conformation (40).Chandler and Pritchard (9) observed an in-

crease in lac gene output (alac, P-galactosidaseper total protein, or, similarly, P-galactosidaseper OD4w unit), after an increase in replicationvelocity, in agreement with our data (Table 4).Since this increase occurred in the absence of anincreased relative gene dosage (lac genes pergenome), they concluded that relative gene dos-age is not a factor in determining constitutivegene expression. On the other hand, since the

VOL. 150, 1982

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

580 CHURCHWARD, BREMER, AND YOUNG

concentration of lac genes (number of lac genesper OD4w unit) increased after a step-up toabout the same extent as alac (Fig. 3), theysuggested that the increase in gene output after astep-up is due to an increase in lac gene concen-tration and concluded that gene expression islimited by the gene concentration, which ap-pears to be at variance with our conclusion thatDNA concentration does not limit the rate oftranscription. It is important to note that thegene expression measured as specific enzymeactivity (ajc) depends not only on the synthesisof lac-mRNA, but also on total mRNA synthe-sis, since ribosomes must compete for differentmRNA molecules. We suggest the use of theterm gene activity, defined as rate of transcrip-tion per given gene (rate of initiation of mRNAchains at the promoter considered) instead of theterm gene output. The transcription rate per(constitutive) lac gene did not change by thestep-up in replication velocity (Table 4).The unaltered lac gene activity after the repli-

cation velocity step-up means that also the con-centration of free (cytoplasmic) RNA polymer-ase (in the conformation that binds to lacpromoters) remains unaltered. The law of massaction would predict a decreased concentrationoffree RNA polymerase by the step-up (3). Onlyin two circumstances would this not be the case.(i) If most transcription would come from partic-ularly active genes near the replication origin,the number of polymerase binding sites (permass) would actually not increase since theconcentration of genes near the origin is notaffected by changes in replication velocity. (b)If, for most regulated genes, an increase in copynumber were compensated by a correspondingrepression, the summed affinities of all promot-ers for RNA polymerase would not be signifi-cantly altered. The first effect, i.e., the cluster-ing of active genes near the origin of replication,may be more important, since only this wouldexplain the difference in lac gene expressionafter changes in initiation control (Table 4).

Effects on gene control. If control factors arenot in great excess over their binding sites on theDNA, the lower DNA concentration in TJK16would be expected to bring about an increasedconcentration of free (not DNA-bound) positiveand negative control factors, and activation inthe case of positive factors or repression of thecorresponding genes in the case of negativefactors. The disproportionate change in consti-tutive lac transcription and DNA concentrationobserved here (Table 4) might be an effect ofsuch changes in gene control.

APPENDIXDe_trminaton of lac gene transcription. Transcrip-

tion of constitutive lac genes was found from the

fractional synthesis of 3-galactosidase and totalmRNA synthesis in the presence ofexcess inducer andcyclic AMP (15). The fractional synthesis, a,ac, isdefined as: a, = (synthesis of lac protein)/(synthesisof total protein) = (synthesis of lac mRNA)/(synthesisof total mRNA) x C. The constant C reflects thedifferences in the functional life and ribosome loadingefficiency of lacZ mRNA relative to average or bulkmRNA; the factor is presumed to be invariable underthe conditions compared.To find the transcription per gene, the quotient is

expanded as follows: oqc - [(lac genes per genome) x(lac mRNA synthesis per lac gene)]/(total mRNAsynthesis per genome). alc and total mRNA synthesisper genome was observed and lac genes per genome(relative gene dosage) can be calculated from thereplication velocity and the map location of the lacgene (1). In the particular case of lac, the locationhalfway between the origin and terminus of replicationresults in about one lac gene per genome at anyreplication velocity (lac genes per genome = 1).

Instead of using DNA as a reference unit (pergenome), one may also use cell mass (per OD4w unit)without changing the equation since the referenceunits cancel in numerator and denominator: alc -

[(lac genes per unit mass) x (lac mRNA synthesis perlac gene)]/(total mRNA synthesis per mass). This formis particularly convenient for comparison of conditionswhere total mRNA synthesis per mass (i.e., the de-nominator on the right side) remains unchanged, suchthat lac mRNA synthesis per lac gene a,oJ(DNA/OD4w).

ACKNOWLEDGMENTSThis work was supported by Public Health Service grant

GM 15142 from the National Institutes of Health.We gratefully acknowledge the help of John Ryals with the

hybridization experiments and the patience and skill of SallyRahn in typing this manuscript.

LITERATURE CITED

1. B nn, B. J., K. B. Low, and A. L. Taylor. 1976.Recalibrated linkage map of Escherichia coli K-12. Bac-teriol. Rev. 40:116-167.

2. Bremer, H., and G. Churchward. 1977. An examination ofthe Cooper-Helmstetter theory ofDNA replication and itsunderlying assumptions. J. Theor. Biol. 69:645-654.

3. Brewer, H., and D. G. Dalbow. 1975. Regulatory state ofribosomal genes and physiological changes in the concen-tration of free ribonucleic acid polymerase in Escherichiacoli. Biochem. J. 150:9-12.

4. Bremer, H., R. Young, and G. Churchward. 1977. Initia-tion and termination ofDNA replication in bacteria after astepwise increase in the velocity of replication. J. Bacte-riol. 130:92-99.

5. Brwr, H., and D. Yua. 1968. RNA chain growth rate inEacherichia coil. J. Mol. Biol. 38:163-180.

6. Br-unchede, H., and H. Bremer. 1976. Synthesis of phageand host specific RNA in Escherichia coil infected with afragment of bacteriophage TS. Virology 75:355-367.

7. Brunachede, H., T. L. Dove, and H. Brewer. 1977. Estab-lishment of exponential growth after a nutritional shift-upin Escherichia coll B/r. J. Bacteriol. 129:1020-1033.

8. Cha_aern, M. J. 1976. Interaction of RNA polymersewith the DNA template, p. 159-191. In R. Losick and M.Chamberlin (ed.), RNA Polymerase. Cold Spring HarborLaboratory, Cold Spring Harbor, N.Y.

9. Chadler, M., and R. Prtchard. 1975. The effect of geneconcentration and relative gene dosage on gene output inEscherilchia coli. J. Bacteriol. 138:127-141.

J. BACTERIOL.

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

GENOME CONCENTRATION AND GENE ACTIVITY

10. Choung, K., E. Estiva, and H. Bremer. 1981. Genetic andphysiological characterization of a spontaneous mutant ofEscherichia coli B/r with aberrant control ofDNA replica-tion. J. Bacteriol. 145:1239-1248.

11. Churchward, G., and H. Bremer. 1977. Determination ofthe DNA replication time in exponentially growing Esche-richia coli B/r. J. Bacteriol. 130:1206-1213.

12. Churchward, G., E. Estlva, and H. Bremer. 1981. Growthrate-dependent control of initiation of chromosome repli-cation in Escherichia coli. J. Bacteriol. 145:1232-1238.

13. Collins, J., N. P. FWI, P. Jorgenson, and J. D. Friesen.1976. Gene cloning of Escherichia coli chromosomalgenes important in the regulation of ribosomal RNAsynthesis, p. 356-367. In N. C. Kjeldgaard and 0. Maaloe(ed.), Alfred Benzon Symposium IX, Munksgaard, Co-penhagen. Academic Press, Inc., New York.

14. Cooper, S., and C. E. Helnstetter. 1968. Chromosomereplication and the division cycle of E. coli B/r. J. Mol.Biol. 31:519-540.

15. Dalbow, D. G., and H. Bremer. 1975. Metabolic regulationof P-galactosidase synthesis in Escherichia coli. Biochem.J. 150:1-8.

16. Daneo-Moore, L., and G. D. Schockman. 1977. The bacte-rial cell surface in growth and division, p. 653-715. In G.Poste and G. L. Nicholson (ed.), The synthesis, assemblyand turnover of cell surface components. Elsevier-NorthHolland Biomedical Press, Amsterdam.

17. Dennis, P. P. 1972. Regulation of ribosomal and transferribonucleic acid synthesis in Escherichia coli. J. Biol.Chem. 247:2842-2845.

18. Dennis, P. P., and D. H. Nordan. 1976. Characterization ofthe hybridization between purified 16S and 23S ribosomalribonucleic acid and ribosomal deoxyribonucleic acidfrom Escherichia coli. J. Bacteriol. 128:28-34.

19. Ellwood, M., and M. Nomura. 1982. Chromosomal loca-tions of the genes for rRNA in Escherichia coli K-12. J.Bacteriol. 149:458-468.

20. Hansen, F. G., and K. Rasmussen. 1977. Regulation of thednaA product in E. coli. Mol. Gen. Genet. 155:129-225.

21. Helmstetter, C., and S. Cooper. 1968. DNA synthesisduring the division cycle of rapidly growing E. coli B/r. J.Mol. Biol. 31:507-518.

22. Helmstetter, C., and 0. Pleruccl. 1976. DNA synthesisduring the division cycle of three substrains of Escherich-ia coli B/r. J. Mol. Biol. 102:477-486.

23. Helmstetter, C. E. 1967. Rate ofDNA synthesis during thedivision cycle of Escherichia coli B/r. J. Mol. Biol.24:417-427.

24. Iwakura, Y., K. Ito, and A. Ishihama. 1974. Biosynthesisof RNA polymerase in Escherichia coli. I. Control ofRNA polymerase content at various growth rates. Mol.Gen. Genet. 133:1-23.

25. Kawakami, K., T. Saltoh, and A. Ishlhama. 1979. Biosyn-thesis of RNA polymerase in Escherichia coli. IX.Growth-dependent variation in the synthesis rate, con-

tent, and distribution of RNA polymerase. Mol. Gen.Genet. 174:107-116.

26. Kennel, D. 1968. Titration of the gene sites on DNA byDNA-RNA hybridization. II. The Escherichia coli chro-mosome. J. Mol. Biol. 34:85-103.

27. Koch, A. 1970. Overall controls of the biosynthesis ofribosomes in growing bacteria. J. Theor. Biol. 28:203-231.

28. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the Folin phenolreagent. J. Biol. Chem. 193:265-275.

29. Maal#e, 0. 1969. An analysis of bacterial growth. Devel-op. Biol. Suppl. 3:33-58.

30. Maal#e, 0. 1979. Regulation of the protein-synthesizingmachinery-ribosomes, tRNA, factors and so on, p. 487-542. In R. F. Goldberger (ed.), Biological regulation anddevelopment, vol. 1, Gene expression. Plenum PublishingCorp., New York.

31. Miller, J. T. 1972. Experiments in molecular genetics.Cold Spring Harbor Laboratory, Cold Spring Harbor,N.Y.

32. Nerlich, D. P. 1972. Regulation of ribonucleic acid syn-thesis in growing bacterial cells. I. Control over the totalrate of RNA synthesis. J. Mol. Biol. 72:751-764.

33. Norris, T., and A. Koch. 1972. Effect of growth rate on therelative rates of synthesis of messenger, ribosomal, andtransfer RNA in Escherichia coli. J. Mol. Biol. 64:633-650.

34. Pritchard, R. H., and A. Zarltsky. 1970. Effect of thymineconcentration on the replication velocity of DNA in athymineless mutant of Escherichia coli. Nature (London)226:126-131.

35. Rose, T. K., and C. Yanofsky. 1972. Metabolic regulationof the tryptophan operon of Escherichia coli: repressor-independent regulation of transcription initiation frequen-cy. J. Mol. Biol. 69:103-108.

36. Ryals, J., and H. Bremer. 1982. reUA-dependent RNApolymerase activity in Escherichia coli. J. Bacteriol.150:168-179.

37. Scaife, J. 1976. Bacterial RNA polymerases: the geneticsand control of their synthesis, p. 207-225. In R. Losickand M. Chamberlin (ed.), RNA polymerase. Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.

38. Seeburg, P. H., C. Naulein, and H. Schaller. 1977. Inter-action ofRNA polymerase with promoters from phage fd.Eur. J. Biochem. 74:107-113.

39. Shepherd, N., G. Churchward, and H. Bremer. 1980.Synthesis and activity of ribonucleic acid polymerase inEscherichia coli. J. Bacteriol. 141:1098-1108.

40. Travers, A., and R. Buckland. 1973. Heterogeneity of E.coli RNA polymerase. Nature (London) New Biol.243:257-260.

41. Yuan, D., and V. Shen. 1975. Stability of ribosomal andtransfer ribonucleic acid in Escherichia coli B/r aftertreatment with ethylene-dinitrilotetraacetic acid and ri-fampicin. J. Bacteriol. 122:425-432.

VOL. 150, 1982 581

on March 3, 2021 by guest

http://jb.asm.org/

Dow

nloaded from