FIS-Dependent trans Activation of Stable RNA Operons of Escherichia coli under Various Growth Conditions

HAL Id: hal-01898766

https://hal.archives-ouvertes.fr/hal-01898766

Submitted on 18 Oct 2018

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

FIS-Dependent trans Activation of Stable RNA Operons

of Escherichia coli under Various Growth Conditions

Lars Nilsson, Hans Verbeek, Erik Vijgenboom, Cornelis van Drunen, Anne

Vanet, Leendert Bosch

To cite this version:

Lars Nilsson, Hans Verbeek, Erik Vijgenboom, Cornelis van Drunen, Anne Vanet, et al.. FIS-Dependent trans Activation of Stable RNA Operons of Escherichia coli under Various Growth Con-ditions. Journal of Bacteriology, American Society for Microbiology, 1992, 174 (3), pp.921-929. �10.1128/jb.174.3.921-929.1992�. �hal-01898766�

0021-9193/92/030921-09$02.00/0

Copyright© 1992, American Society for Microbiology

FIS-Dependent

trans

Activation of Stable

RNA

Operons

of

Escherichia coli under Various

Growth Conditions

LARS NILSSON,t HANS VERBEEK, ERIKVIJGENBOOM, CORNELISVANDRUNEN,

ANNE VANET,t AND LEENDERT BOSCH*

Departmentof Biochemistry, Leiden University, Gorlaeus Laboratories, P.O. Box9502,

2300 RA Leiden, TheNetherlands Received 14 June1991/Accepted 14 November 1991

InEscherichia colitranscription of the tRNA operonthrU(tuJB)andthe rRNA operon rrnB istrans-activated by the protein FIS. This protein, which stimulates the inversion of various viral DNA segments, binds

specifically to a cis-acting sequence (designated UAS) upstreamof the promoter of thrU (tuJB) and the P1 promoter of the rrnB operon. There are indications that this type of regulation is representative for the regulation of more stable RNA operons. In the present investigation we have studied UAS-dependent transcription activationofthe thrU(tuiB)operonin the presenceand absence ofFISduringanormal bacterial growthcycle andafter anutritional shift-up. Inearly log phase theexpressionof the operon risessteeply in

wild-typecells, whereafter it declines. Concomitantly, a peakof the cellular FIS concentration is observed.

Cellsin thestationary phasearedepleted of FIS. The rather abrupt increase of transcription activation depends

onthenutritionalqualityof the medium. It isnot seenin minimalmedium. Afterashift from minimaltorich medium, a peak of transcription activation andof FIS concentration is measured. Thispeakgetshigherasthe medium gets more stronglyenriched.We conclude thatacorrelation between changes of theUAS-dependent activation of the thrU (tuJB) operon and changes of the cellular FIS concentration under a variety of

experimental conditions exists. This correlation strongly suggests that the production of FIS responds to

environmentalsignals, thereby trans-activating the operon. Cells unabletoproduce FIS(fiscells) also showan

increaseofoperontranscriptionin theearly log phase andafteranutritionalshift-up,albeit lesspronounced

than that of wild-type cells. Presumably it is controlled by the ribosome feedback regulatory system. cis activationofthe operon by the upstreamactivatorsequenceis apparent in the absenceofFIS. This activation is constant throughout the entire growth cycle and is independent of nutritional factors. The well-known growth rate-dependent control, displayed by exponentially growing cells studied under various nutritional

conditions, is governed by two regulatory mechanisms: repression, presumably by ribosome feedback inhibition, andstimulation by transactivation. FISallows veryfastbacterial growth.

Thesynthesisof rRNA of Escherichia coli is finely tuned

to the cell's environmental conditions. Cells growing in a constantenvironmentdonotshowasignificantturnoveror a significant buildup of free rRNA or vacant ribosomes,

exceptatverylowgrowthrates(for reviews,seereferences 20, 21, and 26). Consequently, ribosomes are utilized at

maximal ornear-maximal capacity. Upon alteration of the

nutritionalcapacity of the medium, leadingto adifferent but

constantenvironment, cellspromptly readjust the synthesis of their rRNA and tRNAto meetthe demands ofanaltered

growth rate. In exponentially growing cells the

concentra-tionof ribosomes (and of rRNA) thusappears tobe

propor-tionaltogrowthrate(6, 8, 9). The mechanismunderlying this so-calledgrowthrate-dependentcontrol has been described

as a feedback inhibition of rRNA synthesis by ribosomes

(ribosomefeedback control) (5, 12, 13, 29)and/orinhibition

by ppGpp (stringent control) (4), the concentration of this

nucleotidebeingafunction ofthe growth medium (28). Gaal andGoursereported thatE. colimutantsunable to produce

ppGppare still undergrowth rate-dependent control. They concluded thatppGpp isnottheonly factor involved in this

type of regulation (7).

*Correspondingauthor.

tPresent address: Department of Cell Biology, University of Stockholm, S-106 91 Stockholm, Sweden.

t Present address: Institut deBiologie Physico-Chimique, 75005

Paris, France.

Interestingly, it has been reported that the synthesis of tRNA issubjectto the same regulatorymechanisms as the synthesis of rRNA (14, 32). While studying the regulation of thetRNA operonthrU

(tuJB),

werecently showed that it is also regulated by a positively acting control system.

Up-streamof thisoperonacis-actingsequenceisfound, deletion of which results in an80to 90%drop oftranscription (31). Thissequence,called UAS forupstreamactivatorsequence,

appearedto be the targetofatrans-activating protein (34), whichwe subsequentlyidentifiedasthe protein FIS (3, 25).

Up till then this heat-stable protein was only known tobe involved in site-specific recombination (2, 17, 30). It stimu-lates theinversion of various viral DNAsegmentsbybinding

to arecombinational enhancer (16, 19).

It may be envisaged that more stable RNA operons are

activated intransby thissystem. Accordingly, the UASs of the tyrT, metY, thrU(tufB), and rrnB operonsall bindone

and thesameprotein (25, 33). Sequencesupstreamof theP1

promoters of all rRNA operons and the promoters of 13

tRNA operons (but not all tRNAoperons) match the con-sensus sequence for FIS-binding sites (18, 33). Recently

Rossetal.(18, 27)independently demonstratedthat FISacts asthetransactivator ofthe rrnB operon. Inthiscontextthe

question of under which conditions the trans activation controlsystembecomes operativearises.Preliminary stud-ies fromthislaboratory (3,25) revealedlargefluctuations in

the trans activation of the thrU (tufB) operon during a

normal bacterialgrowth cycle. Here wehave studied these

922 NILSSON ET AL.

TABLE 1. E.coli strains used inthisinvestigation

Strain Genotype or phenotype Reference orsource

MC1000 lacX74araDl39(araleu)7697 galU galKStrr 15

MC1000-fis767 lacX74 araD139(araleu)7697 galUgalKStrrfis Kanr 15

JM101 supEthi(lac-proAB) 22

JM101-fis767 supE thi (lac-proAB) fis Kanr Ourlaboratory

fluctuations in moredetail. We alsoaskedwhether changes

inthelevel oftransactivationoccurinresponseto environ-mental signals and whether they are accompanied by changes in thecellular FIS concentration. Of further interest was whether growth rate-dependent control is solely gov-erned by ribosome feedback (see above) orwhether trans

activation also plays a role in this process. Finally we describe the benefits for thecellof having FIS.

MATERIALS ANDMETHODS

Strains, plasmids, and growth media. TheE. coli strains used inthisstudyarelisted in Table1.Theplasmid pDS10 is described in reference31. Itharborstheoperonfusion thrU

(tufB)':galK, and the UAS extends from position-176tothe

promoter. pDS10AUAS is identicalto pDS10exceptthat it carriesadeletion extending from -500to-57.

M9 minimalmedium was supplemented with thiamine (1

jig/ml), essential amino acids (20 ,ug/ml), and succinate (1.0%) (30);inthecaseof minimal medium plus amino acids

supplementation was with 0.5% casein hydrolysate. LB

medium(per liter: 10gof Bacto Tryptone,5gof BactoYeast

Extract, 10gofNaCl, pH 7.5)wasprepared by the method

of Miller (23). In the nutritional shift-up experiment 0.1 volume ofa 1Ox concentrated LB medium without NaCl was added. Brain heart infusion medium was prepared

accordingtothe manufacturer'sdescription (Difco Labora-tories, Detroit, Mich.). In most of the experiments with a rich medium the mediumwasreplaced by LBplus glucose

(1%), since this improved the reproducibility oftheresults.

Determinationofgalactokinase activity.The bacteria

trans-formedwithpDS10 orpDSAUAS weregrownovernightin

the medium indicated. The cultures werediluted and

incu-bated further at 37°C. At the times indicated the optical densityat600nm

(ODwo)

was measured and samples were

withdrawn. Galactokinase activities and plasmidcopy

num-bersweredeterminedbythemethod of Adams and Hatfield

(1) with modifications byvanDelftetal. (31).

Bacterial growth.Growthwasmeasuredby reading

OD6w

(inmostcases)orby determinationof thedry weightof the

biomass(see Fig. 8).Inthe lattercasecellsweresedimented

andwashedwith0.83%NaCl,whereafter the sedimentwas

weighed afterlyophiization.

Determination ofthecellular FISconcentration. For each

determination cellextractswereprepared fromtwo

indepen-dentbacterial cultures (40hin minimalmediumpriortothe

shift-upandovernightpriortoagrowth cycleinLBmedium

plus 1% glucose). After sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and Western

immunoblotting with FIS antibodies pretreated with an extract ofthefis strain MC1000-fis767, FIS was detected

with the ECL fluorescence kit of Amersham. Fluorescence signals were quantified by determining peak areas with a

laserscanner.Theseareas werecorrelated with theareaofa cross-reactingprotein representativefortheprotein

concen-tration inthe samples. Therelative FIS concentration

plot-ted in thefigures represents the indirect FIS concentration/ total soluble protein. It is corrected for blot and detection

efficiencybyusing purified FIS as aninternalstandard. The standard deviation was calculated for each point with the data of three independentWestern blots and two scans of each blot. Thelargest deviation found was 15%.

RESULTS

Experimental strategy. In order to study theexpression of thethrU(tufB)operon invivo under various cellular growth conditions we used the plasmid-borne operon fusion as previously described (31). This fusion, thrU

(tuJB)':galK,

which puts theexpression of galK under control of the thrU

(tuJB)promoter, was introduced intogalK-defective E. coli strains. Expression of the operon was studied by measuring

galactokinaseactivity/femtomoleofplasmid. Inherent to this procedure is that a difference in life span of the galactokinase protein and the transcript may lead to anoverestimationof thenumber oftranscripts present at the times indicated and thatgalactokinaseactivitieslagsomewhatbehind changes in FISconcentration andtranscription. Whenrelevant, this is

pointedoutin thetext(e.g.,seeFig. 3A). Large fluctuations in galactokinase activity are observed during a bacterial growthcycle and after a nutritionalshift-up (cf.Fig. 2 to 4). These are duetosynthesisanddegradationof the

galactoki-nase protein, since they cannot be ascribed to comparable

fluctuations inplasmidcopy numbers.Changes in therateof

galactokinase mRNA or protein degradation, occurring

un-der the variousgrowth conditions, may thus bias the results. This drawback can beovercome asindicated below.

Since transcription activation of the operon depends on

the presence of the UAS, galactokinase activities were

measured in cells carrying a plasmid with an intact UAS

(designatedUAS+cells)orwith theUAS deleted

(designat-ed UAS- cells). Other regulatory mechanisms, known to

control stable RNA synthesis such as ribosome feedback inhibition andstringentresponse, have DNAdeterminantsin theregionfromposition -50to +1(11).DNAdeterminants oftranscriptionactivation, however,arefound in the

region

extendingfrom -131to -48(25, 31, 34). Thisenabledus to

distinguishbetweentheeffectofUAS-dependentactivation and thatofrepression by anotherregulatory mechanism(s).

Galactokinase activities of UAS+ cells thus reflect the effects of all regulatory mechanisms including activation,

whereas theactivities of UAS- cells reflect theeffectsof all mechanisms except activation. The ratio of galactokinase

activities of UAS+ and UAS- cells isa measure of activa-tion. It isindependent ofchanges in the rateof mRNA or

protein degradation.

We have alsostudiedtheeffect of FISonactivation andon

other regulatory systems by performing the same

experi-mentsinastrainlackingafunctionalfisgene (fis cells).The

complete lackof FIS infis cellswasconfirmed byWestern blotting withFIS-specific antibodies(datanotshown).

Expression of the thrU

(tall)

operon during a bacterial

E E 400 * 200 :R A cm time(min.) 20 0 10~~~~~~~~~ 02 0.1 0 0.05 0 100 200 time(min.)

FIG. 1. Expression of the plasmid-borne thrU (tufB) operon

during abacterial growth cycle. E. coli MC1000 was transformed

with pDS10 or the plasmid deletion derivative lacking the UAS, pDS10AUAS (see Materials and Methods). The cells were grown

overnight in brain heart infusionmedium,dilutedtoanOD600of0.1, and incubated at 37°C. (A) Galactokinase activities expressed in

unitsperfemtomole of plasmidperhour.0,UAS+ cells;A,

UAS-cells; O,

OD6w.

(B) Ratios of galactokinaseactivities inUAS+and

UAS- cells (0) as calculated from the data in panel A. Standard

errors(SE) of the ratiosweredeterminedby thefollowingformula:

[(SEof UAS+ activity/UAS+ activity)2 + (SE of UAS- activity/ UAS-activity)2]05.The errorbarsrepresenttheSE, ineachcase

based on three independent experiments. O,

OD6..

Background

levels ofgalactokinase activities, representing nonspecific binding of radioactive materialtothe membraneandenzymaticactivityat0°C,

were7% ± 2% for UAS+ cells and 6% ±4% forUAS- cells.

growth cycle. With the strategy outlined above the

expres-sion of the thrU (tufB) operon was investigated during a

normal bacterial growth cycle in various media. Overnight

cultureswere diluted to an

OD6.

of0.1, and galactokinase

activitiesweremeasured atvariousstagesthereafter. Figure

1A shows the results obtained with cells grown in a rich

medium (brainheart infusion). Large fluctuations in galac-tokinase levelsareseenduringthe growth cycle, which is in

agreement with our previous results (25). Almost

immedi-ately after dilution ofa stationary phase culture in fresh

medium, alargeincrease ofgalactokinaseactivityoccursin UAS+cells. Thatthisincrease ismainly duetoactivation of transcription can be concluded from the behavior of

UAS-cells.The latter cells also showanincrease ingalactokinase

activity, but it is much less pronounced. Apparently, an

additional regulatory mechanism differing from activation

becomes operative afterreinitiation ofgrowthbut the

stim-ulation of transcription by this mechanism is greatly

out-weighed by the activation of the operon. The increase in

activation is clearly reflected by the ratio ofgalactokinase

activities in UAS+ and UAS- cells (Fig. 1B).

Theinitial rise ingalactokinase activities is followedbya

dropinboth UAS+ and UAS- cells. The decline in

activa-tion is reflected bya decrease ofthe ratio ofgalactokinase

activities in UAS+ and UAS- cellsfrom 17toapproximately

10(Fig. 1B).Afterthis rather steepdrop,activation declines

further,

albeit more slowly, asindicated by ratios of 4 to 5

when cells approach the stationaryphase.

Sensingthe nutritionalquality ofthemedium.Theincrease inactivation inearlylog phaseisaffectedbythecomposition

of the medium. This became apparent by studying the

UAS+-to-UAS- ratios during outgrowth of overnight

cul-tures in various media. The ratio varies from less than 5 in

minimal mediumto approximately 16in LB medium and 17

in brain heart infusion medium (maximal ratios are not

always

observed at the same cell concentration). Even

though we cannot fully rule out that we have missed the

exact activation peak during the growth cycle, the large

differences observed indicate that the cells sense the

nutri-tional quality of the medium and respond with an altered

activation level.

Transcription regulationinthe presence andabsence ofFIS.

Transcription activation of the thrU (tufB) operonis depen-dent on the protein FIS (3, 25), as is clearly shown by

comparingthe expression ofthe operon inwild-type andfis

cells (Fig. 2). Overnight cultures ofboth types of cells are

devoid of FIS (see below). Accordingly, deletionof theUAS

has the sameeffectonthegalactokinaseactivities of station-arywild-type andfis cells, i.e., an approximately threefold reduction of the galactokinase activities (compare zero time

activities inFig. 2A andB). Apparently,we aredealinghere

witha cis effect ofthe UAS in the absence of FIS. Thiscis

effect should beconstantthroughoutthe entire growthcycle

infis

cells.This is indeed what theexperiment shows, ascan

be concluded from the UAS+-to-UAS- ratios (Fig. 2B and

Table 2). A FIS-independent effect of the UAS on the

expression of therrnB operonhas beenshown, both in vivo

and in vitro, by Gourse and coworkers (27). In contrast to

what is seen in fis cells, the UAS+-to-UAS- ratio rapidly increases in wild-type cells (Fig. 2A and Table 2). The

activation that becomes operative in these cells thusis dueto

twoeffects: one is dependentonFIS occurringimmediately after reinitiation ofgrowth, and the other most likely is a

FIS-independent cis effect induced by the nucleotide se-quence, since it acts continuously and at a constant level.

Figure 2A and B further show expression of the thrU

(tufi) operonafter deletion ofthe UAS in wild-type andfis cells. As mentioned above, an additional control

mecha-nism, different from UAS-dependent activation, becomes

apparent here. We suggest that it reflects derepression of ribosome feedback inhibition control.

Interestingly, the UAS-independent transcription persists

foralonger period of time infiscells than in wild-typecells,

leading to asignificantly highergalactokinase level infis cells than in wild-type cells. A similar phenomenon is observed

with therrnBoperon (27). Although further experiments are needed, these data suggest that UAS-dependent activation and theadditional regulatory system affect each other.

Ascanbe seen in Table 2 and Fig. 2A and B, expression of the thrU (tufB) operon with an intactUAS isalmostequal

924 NILSSON ET AL. X200 E

a

cL i100 C co time(min.) tme(min.)

FIG. 2. Effectsof FIS and UAS on thrU(tuiB)expression during

growthcycle. Wild-typecells (MC1000) (A) andfiscells

(MC1000-fis767) (B) were transformed with the plasmids described in the legend to Fig. 1.Overnightcultures werediluted to anOD6 of 0.2 in LB medium and grown at 37°C. Galactokinase activities are

expressedasdescribedin the legend toFig. 1. Forsymbol

defini-tions,see thelegend toFig. 1.

inwild-typeandfis cells foracertainperiodof time (between

OD6.

of0.8and 1.6). Sincefis cells cannot use the

tran-scription activationsystem, theyapparentlycompensatefor it by derepressing the ribosome feedback control system.

Cellularlevels of FIS. Thelarge fluctuations in

transcrip-TABLE 2. Galactokinaseactivity ratios of various cell types

(MC1000andMC1000-fis767) determinedduring

thegrowthcyclea

Galactokinaseactivityratio of:

OD6'WTUAS+/ WTUAS+/ fisUAS+/

WTUAS- fis UAS+ fis

UAS-0.2 4.1 ± 0.87 1.5 ±0.52 2.9± 1.0 0.4 12.0± 2.9 1.6 ±0.26 3.3± 0.45 0.6 6.7 ± 2.4 2.3 ±0.91 3.6± 1.2 0.8 6.5 ± 1.2 0.84±0.22 3.6± 1.3 1.6 6.8 ± 1.3 0.86± 0.16 3.0± 0.2 2.2 2.5± 0.32 0.59± 0.15 2.6 ±0.55 2.4 3.5 ±0.67 0.56±0.12 3.8 ± 1.0

a SeelegendstoFig.2Aforwild-type (WT)cellsandFig.2Bforfiscells.

c 0 co C3

1

1

Co 4 0.2 2 0.1 0 0.05 0 100 200 300 time(min.)

B

0.8 ii

~~~~~~~~~~~~~~~~~~~~~0.5

time(min.)

FIG. 3. Cellular FIS concentration during growth cycle. Over-night cultures of the wild-type strain MC1000 were diluted to an

ODC of0.15 and grown at37°Cin LB medium plus 1% glucose (A) or in minimal medium (1% succinate) (B). At the times indicated the cellular FIS concentration was determined asdescrbed in Materials and Methods. *, relative FIS concentration (see Materials and

Methods);O.

OD6w.

tion activation of the thrU(tufB) operon duringthegrowth cycleraise thequestionof whether the cellularlevel of FIS also varies. Thompson et al., using a FIS-DNA binding

assay, reported that the FIS level dropped 70-foldas cells

wentfrom late logtostationary phase (30). Wedetermined the cellular FISconcentrationduring the entire growth cycle

by usingWesternblotting. Cells cultured inarich medium (LBplus1%glucose)showarapidincrease of their FIS level in the early log phase (Fig. 3A). A peak value of FIS is reachedapproximately 75 minafterinitiation of thegrowth cycle, whereafter a rather steep decline sets in. In the stationary phase FIS has dropped toa level atwhich it is undetectable andsohastranscriptionactivation. The

maxi-mum level of FIS does not exactly coincide with that of

galactokinase activity (cf. Fig. 2A and 3A). This is to be expected, since the galactokinaseactivities giveanindirect estimate of the number of transcripts and will lag behind changes in transcription. Conceivably, the fluctuations in FIS level largely govern the fluctuations in transcription

activationduringthegrowth cycleinthis medium.

Cells growing in minimal medium(1% succinate) do not

show an abrupt rise in FIS content. The slight gradual

increaseof FIS observed in

Fig.

3B isatthe lower limits of detection and may be of

questionable significance.

We conclude that the fluctuations in the cellular concentration of FIS are sensitivetoenvironmental

signals.

Since

transcrip-tion activatranscrip-tion also varies with

changes

in thecompositionof the

medium,

a correlation between

changes

in the

UAS-dependent transcription

activation of the thrU

(tufil)

operon and

changes

of the cellular FIS concentration becomes apparent. This correlation

strongly

suggests that environ-mental conditions

signal

trans activation of the operon by

FIS. Furtherconfirmation of this correlation was obtained

by studying

the effectofa nutritional

shift-up.

Effects of a nutritional

shift-up.

Figure

4 shows the

re-sponseof

exponentially growing wild-type

andfis

cellsto a

shift from minimal mediumtoLB medium. The

growth

rate

of

wild-type

cells

promptly

increases from 0.4 to 2.2

dou-blings

per h

(not

shown),

while

galactokinase activity

in-creases

approximately

10-foldover a

period

of2hafter the shift

(Fig.

4A).

This enhanced

expression

of the thrU

(tuJB)

operon is

greatly

due to a rise in

transcription activation,

since the

UAS+-to-UAS-

ratio increases

approximately

fourfold

during

this

period.

Thecells also

respond

witha

change

in their FIScontent

(Fig.

4B).

As illustrated above

(Fig. 3B),

cells

growing

in

minimal medium do not

display

sudden fluctuations in FIS

concentration

during

the

growth cycle,

and this

concentra-tion remains very low.

Immediately

after

changing

the

composition

of the

medium,

however,

a steep rise of FIS

occurswithina1-h

period,

whereafter it also declines rather

steeply

and levelsoffata

relatively

elevated level.

Cells unable to

produce

FIS

(fis

cells

[Fig.

4C]) likewise

sense this rather drastic

change

in the

composition

of the medium. The

galactokinase activity rises, although

not tothe

same level as in

wild-type

cells,

but the

UAS+-to-UAS-ratio remains constant.

Apparently,

a

regulatory

mecha-nism,

different fromtrans

activation,

is

responsible

for this enhancement of the thrU

(tuiB)

expression.

A limited nutritional shift up from 1% succinate to 1%

CasaminoAcids

plus glucose

also enhances the

growth

rate

of

wild-type

cells

(from

0.4 to 1.1

doublings

per h

[not

shown])

and the

expression

ofthethrU

(turB)

operon,as can

beconcluded from

Fig.

5.Galactokinase

activity

risesover a

period

of about 1h

(Fig.

5A).

Thattransactivation

partici-pates in the enhanced

transcription

canbe concluded from

theincreasesinthe

UAS+-to-UAS-

ratio and in cellular FIS level

(Fig.

5A and

B,

respectively).

Finally, fis

cells also

respond

withanincreasein

galactokinase activity (Fig.

5C). Sincetransactivationdoesnottake

place

in these cells

(the

UAS+-to-UAS-

ratioremains

approximately

constant),

the elevated

expression

of theoperon is

governed

by

adifferent control

mechanism,

most

likely

ribosomefeedbackcontrol.

Apparently

the responses of

wild-type

and

fis

cells to a

limited nutritional

shift-up

are

quantitatively

rather similar but

qualitatively

different in the

regulatory

mechanism

re-sponsible

for the

change

in

transcription.

Pulse-chase experiments.

Although FIS-dependent

trans

activation sofar has

only

been demonstrated in vivoandin vitro for the thrU

(turB)

and rrnB operons (3, 18, 25, 27),

various data suggest that the

regulation

of these operons is

representative

for that of more stable RNA operons (see

above).

Inthiscontextonemay ask whether thestimulation ofstable RNA

synthesis by

a nutritional

shift,

as studied

more

directly

by pulse-chase experiments,

differs in

wild-type

andfis

cells. Adirect

comparison

of the results ofFig.

4with those ofa

pulse-chase experiment

inwhich minimal medium is shiftedtoLBmedium is

technically

not

feasible,

C 0 (. c 8 CO) cn LD time(min.)

FIG. 4. Response to an extensive nutritional shift-up. Cells grown overnight in minimal medium were diluted 10-fold with minimal medium and grownat37°Cfor 2h,whereafter(arrow) the mediumwasshifted toLB mediumplus 1%glucose. At thetimes indicatedgalactokinase activityand FIS concentration were deter-mined as described in Materials and Methods. (A) Galactokinase activities ofwild-type MC1000 cells transformed with pDS10 (0)

(UAS+ cells) orwithpDS10AUAS (A) (UAS- cells);(B) relative cellular FIS concentration of MC1000 cells (*); (C) galactokinase

activitiesofMC1000-fis767(fis cells), transformed withpDS10 (0)

926 NILSSON ET AL. 100

time

(min.) 0.6 U)0.4 Er: time(min.) E is

i

L200

M 00 100 0 100 200 time(min.)

FIG. 5. Responsetoalimitednutritional shift-up. Experimental

conditionswereidenticaltothosedescribedinthelegendtoFig. 4,

except that the medium was shifted from minimal to Casamino Acids(1%) plus glucose (1%).

sincethelabelisdilutedoutinLBmedium. Shiftingfrom1%

succinate to 1% Casamino Acids plus 1% glucose was

therefore carried outinstead. Cells receiveda2-min

[3H]u-ridine pulse which was followed by an 8-min chase. The resultsdiffered somewhatdependingonthe pretreatmentof

thecellspriortothe shift, butessentiallytheoutcomewas as

illustratedinFig.6. StableRNAsynthesisinbothwild-type

FIG. 6. Stable RNAsynthesisafter alimited nutritional shift-up, studiedby pulse-chase experiments. MC1000-fis767 (fis)cells were

cultured in minimal medium and shifted fromminimal medium to

Casamino Acids (1%) plus glucose

(1%).

One-milliliter culture

samples were pulse-labeledwith 37 KBqof

[3H]uridine

for 2

min,

which wasfollowedby achasewith0.125mMnonlabeled uridine, whereafter0.2 ml of anonlabeled overnight cultureas acarrierand 3volumesof7%trichloroaceticacid wereadded.Precipitateswere

collected onGF/C filters, washed, dried, andcounted in a

liquid

scintillationcounter

U,

wild-type MC1000cells; O,fiscells.

andfis cells increases immediately after the change of the

medium. We conclude that the response of stable RNA

synthesisto alimitednutritional shiftisrathersimilartothat of thrU

(tufB)

expression. Theresults of Fig. 6 are also in line with those of Fig. 5 in that they do not reveal large

quantitativedifferences in the responses of wild-typeandfis

cells. Qualitatively the responses of both cell types differ;

however, since wild-type cells use trans activationto stim-ulate theexpression of the thrU

(tuJB)

operon(cf. Fig. 5B),

whereasfiscells do not (seeDiscussion).

Growth rate-dependent trans activation. The fact that rrn

operons lacking the UAS are submitted to growth

rate-dependent control(11) indicates that cellsincapableoftrans

activation utilize the mechanismofnegative regulation by ribosome feedback. However, since the present

investiga-tion shows that the composition of the medium has a

pronouncedeffect on theUAS-dependenttransactivation,it doesnotseem verylikely thatgrowthrate-dependent regu-lation is solely

governed

by repression of stable RNA

synthesis. In order to investigate this question we have grownUAS+andUAS-cells indifferent media,permitting

a variation of the growth rates of between 0.9 and 2.0

doublingsperh.

Cells

wereharvestedat an

OD6.

of0.4, and theirgalactokinase activities were determined. In this way

thepresentresults arefullycomparabletothoseobtainedby

Gourseandcoworkers(11), whostudiedthe DNA

determi-nantsofgrowthrate-dependent controlof therrnBoperon. As isapparent fromFig.7, bothUAS+and UAS- cells show

alinearrelationshipbetweenthegalactokinase activitiesand thegrowthrates, inaccordancewith thefindingsfor therrnB

operon (11). In this range ofgrowth rates the increase in

galactokinase

activities is 2.6times for UAS-

cells

but4.6

times forUAS+

cells.

Increasedgrowthratethusleadsto an

increased UAS+-to-UAS- ratio. We conclude that two

controlmechanisms underliegrowthrate-dependent

regula-tion: repression, presumably by ribosome feedback, and

stimulationbytransactivation.

400 E °

1F

CL .5

~~~~~~0

100 _ 0 Cu cm 0 1 2

growth rate(doublJhr.)

FIG. 7. Expression of the thrU (tufB) operon in exponentially growing cells at various growth rates. Transformants described in thelegendtoFig.1weregrownovernightin LBmedium and diluted toanOD600of 0.1 in differentmedia. Growth rates were determined

duringtheearly log phase,andgalactokinase activities were mea-sured at an

ODwo

of 0.4.Background galactokinaselevels were6%

± 1.5%forUAS+cellsand20%o±4%for UAS- cells(cf.legendto

Fig. 1). The media used were minimal medium, minimal medium

supplemented with Casamino Acids, LB medium, LB medium

supplemented withglucose, andbrain heart infusion medium (0, UAS+cells;A,UAS-cells).

FISallows very fast cellular growth. We have seen thatthe

expressionof thefis gene is verypronouncedwhen station-ary cells reinitiategrowthinarich medium (Fig. 1 to 3). A

rapid increase of FIS is accompanied by a rise of

FIS-dependent trans activation of the thrU(tufB) operon. One maytherefore expectwild-type cells to grow faster thanfis

cells under these conditions. Thisexpectation isborne out

bythefollowing experiment. Wild-type(MC1000) cells and

fis (MC1000-fis767) cellsweregrown in LBmedium supple-mented with 1% glucose to an

OD6.

of 0.3. The FIS concentration ofwild-type cells has then reached its

maxi-mum (cf. Fig. 3A). At this stage both types of cells were diluted 40-fold, whereafter the

OD6.

was carefully moni-tored. Atregular intervals samples were collected to deter-minethedry weight ofthe accumulated cell mass. As can be concludedfromFig.8,growthratesof 2.2 and 1.4doublings

per hwerefound forwild-type andfiscells, respectively, on the basis ofreadings of

ODwo,

and growth rates of 1.8 and 1.3doublingsper h onthe basisof biomassassays. Theeffect ofthefisdeletion isnotstrainspecific, sinceweobtained the

sameresultwith the strain JM101.

Recently, Gille et al. (10) reported that FIS binds and

bendsthe origin ofchromosomal DNAreplication, oriC,of

E. coli and that FIS is requiredforminichromosome

repli-cation. The difference in growth rate ofwild-type andfis cells, as measured here by reading OD and by assaying biomass, cannot be ascribed to a reduced replication ability. Weconclude that exponentially growingcellsable to trans-activate stableRNA synthesishave anadvantage over cells unabletotrans-activate and that one of the functions of FIS is to allow fast cellular growth (see Discussion).

DISCUSSION

Acceleration of cellular growth is accompanied by an elevated transcription of the thrU (tufB) operon. Such an

U - ~1.000 , 0,2 500/ 0 A. 0.1 200 0,05 100 0,02 50 2 4 time (hrs.)

FIG. 8. Growth of wild-type cells (MC1000) and fis cells

(MC1000-fis767) in rich medium. Cellsgrown toan

OD6w

of 0.3in

LBmedium plus 1%glucosewerediluted 40-fold andgrownat37°C

in the samemedium.Atthetimesindicated,

OD6w

(opensymbols)

was read and the dry weightof the accumulated cell mass(closed

symbols) was determined (see Materials andMethods). O and *, wild-type cells; A andA,fiscells.

acceleration clearly occurs after dilution of an overnight

culture in fresh medium(Fig. 1and2)and afterashift from minimal to rich medium (Fig. 4). In both cases the rise in

transcription is dependent on the UAS, the target of the protein FIS. Deletion of the UAS abolishes the increase of thrU (tufB) expression in the early log phase almost

com-pletely(Fig. 1A). Theratio ofgalactokinaseactivities of cells

carryingtheplasmid-borne operonfusion thrU(tufB)':galK

with an intact or a deletedUAS (the UAS+-to-UAS- ratio)

rises substantially (Fig. 1B). As pointed out above (see "Experimental strategy" above), this ratio is a measure of

transcription activation which is independent ofchangesin the rates of galactokinase mRNA andgalactokinaseprotein

degradation during the course of the experiment. In cells harboring the fis gene, this activation increasesimmediately

aftergrowthacceleration and ingeneraldeclines thereafter.

Incells unable toproduce FIS (fis cells) the

UAS+-to-UAS-ratio remains constant throughout the entire growthcycle.

ThisFIS-independent cisactivationis notaffectedby envi-ronmental conditions. We assume that it is induced by the

nucleotidesequenceof the UAS. Increases of the UAS+-to-UAS- ratioarealsoseeninwild-typecells after anutritional shift-up (Fig. 4A and 5A). In contrast this ratio remains constantinfiscells(Fig. 4C). Concomitant with the changes in transcription activation, the cellular FIS concentration rises in wild-type cells, both in early log phase and aftera

shift-up (Fig. 3A, 4B, and SB). Apparently, a correlation

between changes of the UAS-dependent activation of the thrU(tufB) operon andchanges of the cellular FIS concen-tration exists.

Various lines of evidence indicate that environmental conditions affect the FIS production and the activation of

transcription. First, the UAS+-to-UAS- ratios during bac-terialoutgrowth in freshmedium vary with thecomposition

of the medium. The large rise and fall of FIS, observed during the growth cycle inarich medium (Fig. 3A),are not

observedduringacyclein minimal medium(Fig. 3B). After

a shift from minimal to rich medium, however, both the

cellular FIS content and operon transcription activation increase promptly (Fig. 4B and A, respectively). The

in-928 NILSSON ET AL.

creases of FIS and oftranscription activation are largeras

themedium is more strongly enriched (cf. Fig. 4B and SB).

We conclude that environmental conditions affect

tran-scription activation and the FIS concentration in the same

way. This correlation, apparently occurring underavariety

ofconditions(seeabove), strongly suggests that it is FIS that trans-activatestranscriptionofthe thrU

(tuJB)

operon inE.

coli. Thealternative model, that FIS concentrationschange

in response to changes instable RNAsynthesis,isnot avery logical one and has some rather bizarre implications. FIS-dependent trans activation apparentlyacts as a sensorof the nutritionalquality of the medium. These data raise

interest-ing questions concerning the regulation of de novo FIS

synthesis and the transduction of the environmental signals

involved.

As pointed out in the introduction, anumberof observa-tions indicate that the trans activation of the thrU

(tuJB)

operonisrepresentativefor sucharegulationofmorestable RNA operons.One and the sameproteinbinds in vitrotothe UASs ofthe tyrT, metY, andthrU

(tuJB)

operonsandtothe UAS upstream of the P1promoterof the rrnB operon (25).

FIS-dependent transactivationof the rrnBoperon has been demonstratedbyRoss etal. (18, 27). Sequence comparisons

of upstream regions (18, 33) are also in line with the

assumption that more stable RNA operons (albeit not

nec-essarily all) are submitted to this regulatory system and respond to FIS. If so, this would indicate that one of the basic elements of growth control is the regulation of the de novo synthesis of FIS.

Cells unabletoproduceFIS(fis cells)alsoshow enhanced

transcriptional activityof the thrU(tufB)operon inearlylog phase of thegrowth cycle (Fig. 2B) andafter enrichment of themedium(Fig. 4C andSC),albeit in rich mediato alesser

extent than wild-type cells. Since no

FIS-dependent

trans

activation of the operon takes place in these

cells,

the

UAS+-to-UAS- ratio remains virtually constant,

irrespec-tive of the environmental conditions. The fact that

galactoki-nase activity of UAS+ cells exceeds that of UAS- cells

reflects cis activation of the operon by the UAS in the

absence of FIS.

Significant transcription

of the operon is thus observed infiscellsduringthe

growth cycle

(Fig. 2B).

In fact, wild-type andfis cells only show a

pronounced

difference inUAS-dependent

expression

of theoperon

dur-ingtheinitialtranscriptional jumpin

early log phase

(cf.

Fig.

2A and B). Thereafter, virtually no difference is seen fora

certainperiod oftime (between

OD600

of 0.8 and

1.6).

Apossible explanation of these data is the

following.

In wild-type cells this leveling off at a lower level is due to

ribosome feedback and adrop in FIS concentration.

Ribo-some feedback in these cells may result from a temporary

excess ofribosomes, present after the steep rise in stable RNAsynthesisinearly log phase.

Infis

cells this rise is less

pronouncedandsomay be ribosome feedback inhibition. On the other hand thesecells donot

trans-activate,

so

transcrip-tion in bothcelltypes reachesapproximatelythesamevalue.

Inaccordance with adifferentextentof ribosome feedback in wild-type and fis cells are the results obtained upon

deletingthe UAS of the reporter operon. In

wild-type

cells

expressionof theoperon dropstovery low

levels,

whereas

infiscells theexpressionremainsatthe samelevel forlonger periodsof time (Fig. 2B).

When stable RNA synthesis is studied in

exponentially

growing cellsatvariousgrowthratesby

altering

the compo-sition of the medium, a linear relationship between the

expression of stable RNA operons and the

growth

rate is found(11, 24). Bothribosome feedback andtransactivation

governthis

growth

rate-dependent regulation,

as isevident from Fig. 6, in which the

expression

of the thrU

(tufB)

operonwas

investigated.

Inaccordancewith

previous

stud-iesof therrnBoperon

(24),

growth-rate

dependent

controlis observed when the thrU

(tufB)

operon is studied after deletion of the UAS. From this

finding

one should not

conclude, however, that ribosome feedback control is the

only

mechanism

underlying growth

rate-dependent

regula-tion. The

UAS+-to-UAS-

ratios measuredatvarious

growth

ratesclearlyindicate that both the

positive

and the

negative

control mechanismsare

operative

under these conditions.

Finally

the

question

of whether the bacterial cell benefits from

harboring

afis

gene maybe asked.

First,

therearetwo

counteracting

regulatory

systems; a

positively

operating

trans activation and a

negatively operating

ribosome feed-back may allow fine

tuning

oftRNAand rRNA

synthesis.

A

prerequisite

of such abalanced control is that trans

activa-tion,

like ribosome

feedback, responds

to environmental

signals.

The results of the present

study

demonstrate that this condition is fulfilled.

Second,

FIS enables the cell to

grow very fast. This may be

quite

an

advantage

when cells grow underthe conditions

prevailing

in nature.

Until

recently

theE. coli

protein

FIS was known

only

to

play

a role as a host factor in the

replication

of certain

phages.

As such itdid notseemto contributetothewelfare of the bacterial cell. It has become evidentnowthat cellsdo benefit from this heat-stable

protein

andareabletoswitchon

its

synthesis

when environmental conditions make this de-sirable.

ACKNOWLEDGMENTS

Theplasmids containingthe UASof the thrU

(turB)

operonand deletion derivatives thereofwere generously donated by J. van Delft. Thestrains MC1000 and MC1000-fis767werekindlyprovided

by R. C. Johnson. The technical help of Anneke Kuipers is

gratefully acknowledged.WeareendebtedtoN.Goosen and P.van de Puttefor valuablehelpand suggestions.

TheinvestigationwassupportedinpartbytheCommission of the

EuropeanCommunities, BiotechnologyAction Programme (BAP),

Directorate-GeneralScience,Research andDevelopment,Brussels. L.N.wastherecipientofalong-termEMBOfellowship,and A.V.

wassupported bytheERASMUS program.

REFERENCES

1. Adams,C. W., andG. W.Hatfield. 1984. Effects ofpromoter

strengthsandgrowthconditions oncopynumber of

transcrip-tion-fusionvectors. J. Biol.Chem. 259:7399-7403.

2. Betermier, M., V. Lefrere, C. Koch, R. Alazard, and M.

Chandler. 1989. The Escherichia coli

protein,

Fis:

specific

bindingtotheends ofphageMuDNAandmodulationofphage

growth. Mol. Microbiol. 3:459-468.

3. Bosch, L., L. Nilsson, E.

Vlgenboom,

and H. Verbeek. 1990.

FIS-dependenttrans-activationof tRNA and rRNA operons of Escherichia coli.Biochim.Biophys. Acta1050:293-301.

4. Cashel, M., and K. E. Rudd. 1987.The stringentresponse,p.

1410-1438.In F.C. Neidhardt,J. L.Ingraham, B.

Magasanik,

K. B. Low, and M. Schaechter (ed.), Escherichia coli and Salmonellatyphimurium:cellularandmolecularbiology.

Amer-icanSocietyforMicrobiology,Washington, D.C.

5. Cole, J. R., C. L. Olsson, J. W. B. Hershey, M.

Grunberg-Manago, and M. Nomura. 1987. Feedback

regulation

of rRNA

synthesisinEscherichia coli.J. Mol. Biol. 198:383-392. 6. Dennis, P. P., and H. Bremer. 1974. Differential rate of

ribo-somalprotein synthesis in Escherichia coli B/r. J. Mol. Biol. 84:407-422.

7. Gaal, T., and R. L. Gourse. 1990. Guanosine 3'-diphosphate

5'-diphosphate isnotrequiredforgrowthrate-dependent

con-trolofrRNAsynthesisinE. coli.Proc. Natl. Acad.Sci. USA 87:5533-5537.

8. Gausing, K. 1974.Ribosomal protein in E. coli: rate of synthesis and pool size at different growth rates. Mol. Gen. Genet. 129:61-75.

9. Gausing,K. 1977.Regulation of ribosome production in Esch-erichia coli: synthesis and stability of ribosomal RNA and of ribosomal protein messenger RNA at different growth rates. J. Mol. Biol. 115:335-354.

10. Gille, H., J. B. Egan, A. Rothand, and W. Messer. 1991. The FIS protein binds and bends the origin of chromosomal DNA replication, oriC, of Escherichia coli. Nucleic Acids Res. 15: 4167-4172.

11. Gourse, R. L., H. A. de Boer, and M. Nomura. 1986. DNA determinants of rRNA synthesis in E. coli: growth rate depen-dentregulation, feedback inhibition, upstream activation, anti-termination. Cell 44:197-205.

12. Gourse, R. L., Y. Takebe, R. A. Sharrock, and M. Nomura. 1985. Feedback regulation of rRNA and tRNA synthesis and accumulation of free ribosomes after conditionalexpressionof rRNA genes. Proc. Natl.Acad. Sci. USA 82:1069-1073. 13. Jinks-Robertson, S., R. L. Gourse, and M. Nomura. 1983.

Expression of rRNA and tRNA genes in Escherichia coli: evidence for feedback regulation by products of rRNA operons. Cell 33:865-876.

14. Jinks-Robertson, S., and M. Nomura. 1987. Ribosomes and tRNA, p. 1358-1385. In F. C. Neidhardt, J. L. Ingraham, B. Magasanik, K. B. Low, and M. Schaechter (ed.), Escherichia coli andSalmonella typhimurium: cellular and molecular biol-ogy, vol. 2. American Societyfor Microbiology, Washington,

D.C.

15. Johnson, R.C., C. A. Ball,D.Pferrer, and M.I. Simon. 1988. Isolation of the gene encoding the Hin recombinational

en-hancerbinding protein. Proc. Natl. Acad. Sci. USA 85:3484-3488.

16. Johnson, R. C., and M. I. Simon. 1985. Hin-mediated

site-specific recombination requires two 26 bp recombination sites anda60bp recombinational enhancer. Cell 41:781-791.23. 17. Johnson, R. C., and M. I. Simon. 1987. Enhancers of

site-specificrecombination in bacteria. TrendsGenet. 3:262-267. 18. Josaitis, C. A., T. Gaal, W. Ross, and R. I. Gourse. 1990.

Sequences upstream of the -35 hexamer of rrnB P1 affect promoterstrength and upstream activation. Biochim. Biophys. Acta 1050:307-311.

19. Kahmann, R., F. Rudt, C. Koch, and G. Mertens. 1985. G inversion in bacteriophage Mu DNA is stimulated by a site within the invertase gene and a host factor. Cell 41:771-780. 20. Lindahl, L., and J. M. Zengel. 1982. Expression of ribosomal

genesin bacteria. Adv. Genet. 21:53-121.

21. Lindahl, L., and J. M. Zengel. 1986. Ribosomal genes in Escherichia coli. Annu. Rev. Genet. 20:297-326.

22. Messing, J. 1979. A multipurpose cloning system based on

single-stranded DNAbacteriophage M13. Recomb. DNA Tech. Bull. 2:43.

23. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 24. Miura, A., J. H. Krueger, S. Itoh, H. A. de Boer, and M.

Nomura. 1981. Growth-rate dependentregulation of ribosome synthesis inE.coli:expression of the lacZ and galK genes fused toribosomal promoters. Cell 25:773-782.

25. Nilsson, L., A.Vanet, E.V"genboom, and L. Bosch. 1990. The role of FIS in trans-activation of stable RNA operons of E. coli. EMBOJ. 9:727-734.

26. Nomura, M., R. L. Gourse, and G. Baugham. 1984.Regulation ofthe synthesis of ribosomes and ribosomal components. Annu. Rev.Biochem. 53:75-117.

27. Ross, W., J. F. Thompson, J. T. Newland, and R.L. Gourse. 1990. E.coli Fis protein activates ribosomal RNA transcription in vitro and in vivo. EMBO J. 9:3733-3742.

28. Ryals, J., R.Little, and H. Bremer. 1982.Control of rRNA and tRNA synthesis in Escherichia coli by guanosine tetraphos-phate. J. Bacteriol. 151:1261-1268.

29. Takebe, Y.,A.Miura,D. M.Bedwell,M.Tam,and M.Nomura. 1985. Increasedexpression of ribosomal genes during inhibition of ribosomalassembly in Escherichiacoli. J. Mol. Biol. 184:23-30.

30. Thompson, J. F.,L.MoitosodeVargas, C. Koch, R.Kahmann, and A.Landy.1987. Cellular factorscouplerecombination with growth phase: characterization of a new component in the

site-specific recombinationpathway. Cell 50:901-908.

31. vanDelft, J. H. M., B. Marion,D. S. Schmidt, andL. Bosch. 1987. Transcription of the tRNA-tuJB operon of Escherichia coli:activation,termination and antitermination. Nucleic Acids Res. 15:9515-9530.

32. vanDelft, J. H. M.,H. M.Verbeek, P. J.deJong,D.S.Schmidt, A.Talens,and L. Bosch.1988.Control of the tRNA-tufB operon inEscherichia coli. 1. rRNA gene dosageeffects and growth-rate-dependent regulation. Eur. J. Biochem. 175:355-362. 33. Verbeek, H., L.Nilsson, G. Baliko, and L. Bosch. 1990. Potential

bindingsitesof the trans-activator FISarepresent upstreamof all rRNA operons and of many but not all tRNA operons. Biochim. Biophys. Acta1050:302-306.

34. VUgenboom,E., L.Nilsson, and L. Bosch. 1988. The elongation factor EF-Tu fromE.coli bindstothe upstream activator region of thetRNA-tufBoperon. NucleicAcids Res. 16:10183-10197.