JOURNAL OF BACTERIOLOGY, Sept. 1971, p. 878-888Copyright @ 1971 American Society for Microbiology

Vol. 107, No. 3Printed in U.S.A.

Transport-Limited Growth Rates in a Mutant ofEscherichia coliKASPAR VON MEYENBURG

University Institute of Microbiology, O?ster Farimagsgade 2 A, 1353 Copenhagen K, DenmarkReceived for publication 14 June 1971

Mutants of Escherichia coli B/r have been selected which require increased nu-trient concentrations for half-maximal growth rate. This half-saturation constant(Kn) for growth on glucose is 10-6 and 7 x 10-4 M for the wild type (CP 366) andthe mutant (CP 367), respectively. Similarly, the Km is increased for growth onmany other carbohydrates (20- to 500-fold), for the anions P043 and SO42- (ca.100-fold), and for the uptake of several amino acids (20- to 50-fold). At suffi-ciently high concentrations of the nutrients, mutant and wild type grow equallyfast. The yield in terms of cell mass per milligram of substrate is unaffected by themutation. The phenotype of the parent is reestablished in what appears to be thereversion of a single mutation (kmt) which maps between strA and metB. Thepleiotropic decrease of the affinities for transport of the various nutrients seems tobe the result of a modification of the cell envelope which weakens the attachmentof the various specific binding proteins to the periplasmic membrane. Since themutant Km values are increased considerably, high cell densities can be reached inbatch cultures at growth-rate-limiting substrate concentrations (107 to 108 cells/ml).This allows chemical analysis of the cell composition; the application of the mutantto studies of bacterial physiology as function of growth rate is discussed.

Studies of bacterial populations growing at dif-ferent rates have provided considerable informa-tion about the in vivo states and potentials ofthese cells (21). It is against this backgroundthat we must try to assess the role and signifi-cance of the various biochemical mechanismscurrently being studied in vitro.

At a given temperature, the growth rate can bevaried by choosing media with different carbonor nitrogen sources, or both. In this way, anumber of simple relationships between cellcomposition and growth rate have been observed,which permit the construction of fairly detailedmodels for the control of bacterial growth (20).However, many questions remain open. Thus, wedo not know to what extent the state of agrowing population is determined by the evidentneed for particular enzyme systems for the utili-zation of different carbon sources (e.g., glucose,succinate, acetate, or proline).

Alternatively, the growth rate can be varied byusing chemostatic continuous cultures (25, 30).In this case, the growth rate (,u = doublings perhour) is determined by the effective concentra-tion, S, of a given substrate, which is maintainedas a balance between supply from outside andutilization by the cells [essentially according toMonod's equation (24): tL = ,max SI(Km + S)J.

This technique has not been used as widely asbatch cultures to study the general physiology ofmicrobial growth (6, 41); however, certain com-plex metabolic patterns in yeast have beenstudied (2, 43) which could not have been ana-lyzed thoroughly in batch cultures.The obvious advantage of the chemostat is

that the growth rate can be varied without quali-tatively changing the substrate. However, theideal chemostat requires continuous flow of sub-strate and perfect mixing as well as adequate con-trol of aeration, foaming, and contamination. Theusefulness of the chemostat is limited by thesetechnical prerequisites and by the fact that largesamples cannot be repeatedly withdrawn foranalysis.The approach presented here eliminates most

of these difficulties by offering a practical solu-tion to the problem of establishing batch culturesin which different growth rates are defined by dif-ferent concentrations of one particular substrate.In principle, it has always been possible to do so atextremely low cell densities (24, 30). Toobtain half-maximal growth rate of the commonlaboratory strains of Escherichia coli in a glucoseminimal medium, the glucose concentration mustbe reduced to 10-6 M, i.e., the half-saturationconstant Km for growth on glucose. This implies

878

on April 18, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

TRANSPORT MUTANT OF E. COLI

that the cell densities required for biochemicalstudies (at least 107 to 108 cells/ml) cannot bereached in batch cultures in which the growthrate is limited by the glucose concentration.A mutant of E. coli B/r is described below,

whose Km for growth on various sugars is greatlyincreased. With this strain, cell densities can beobtained, which allow the chemical compositionof the cells to be determined as function of A inthe range from 0.25 to 1.5 doublings/hour, at 37 C.The nature of this pleiotropic mutation is dis-cussed in light of the additional finding that theKm is increased not only for several sugars, butfor some amino acids and anions as well.

MATERIALS AND METHODSOrganism. E. coli B/r thy-, collection number CP

366, was used.Media. Phosphate-buffered AB minimal medium

according to Clark and Maaloe (5) was used; aftersterilization, it was supplemented with various carbonsources at variable concentrations, and with 10 gg ofthymine/ml. Tris(hydroxymethyl)amiroinethane (Tris)minimal medium (18) with 0.2% glucose and 10 Mg ofthymine/ml was used to determine the Km values foruptake of SO42-, PO43, K+, NH +, or Mg2+ byvarying the concentration of the respective ion. Allgrowth experiments were performed at 37 C, exceptwhere noted.

Minimal medium for substrate gradient plates wasas follows: 0.1 g of MgSO,*7H2O, 5 g ofNa2HPO4-2H20, 1.7 g of Na(NH,)HPO4-4H20, 0.75g of KCI; pH adjusted with HCI to 7.0; 1% agar; 10 Mgof thymine/ml; in 1,000 ml.

MacConkey-plates were prepared as follows: in1,000 ml of distilled water, 40 g of MacConkey agarbase (Difco) was added; after autoclaving, sterile glu-cose solution was added.

Measurement of growth and half saturation constantfor growth on different substrates. The optical density(OD) for cultures was determined with a Zeiss PMQII spectrophotometer at 450 nm in I-cm cuvettes. AnOD450 of 0.1 corresponds to 3.3 x 107 cells/ml (9 Mgof protein/ml) of E. coli B/r thy- growing in AB me-dium with 0.2% glucose (doubling time, 42 min). Atlow cell densities (e.g., in experiments with low glucoseconcentrations), 2-cm cuvettes were used, but ODvalues are always given as absorbancy per I cm. Cellnumber and cell size distribution were determined witha modified Coulter counter (38). The growth rate A isexpressed in doublings per hour, and the doublingtimes g in minutes (specific growth rate K = Mi In2).From the growth rates at different substrate concentra-tions, S, the Km for growth was estimated graphicallyby plotting uM 1 against S-1 (see Fig. 2B).

Determination of the half saturation constant foranuno acid uptake. Different radioactive amino acids(14C- or 3H-labeled; 2 to 5 mCi/mmole) were added atdifferent concentrations to cultures growing in ABmedium with 0.2% glucose. The rate of incorporationof radioactivity into protein was determined by sam-pling (0.2 ml) at intervals into 5% trichloroacetic acid,collecting the precipitate on membrane filters, and

counting the radioactivity in a Nuclear-Chicago scintil-lation spectrometer (1). From the rates of incorpora-tion (V, assumed to reflect the rate of amino acid up-take), the Km for amino acid uptake was estimatedgraphically by plotting V- I against (amino acid con-centration) 1; see Fig. 4C.

Determintion of the cell composition. Protein andribonucleic acid (RNA) were determined by a modifi-cation of the procedure described by Koch and Deppe(17). Portions (9 ml) of cultures with OD45, between0.3 and 0.6 are sampled into 2 ml of ice-cold 1.2 N per-chloric acid (PCA) and centrifuged after 20 min; thepellet is resuspended in 0.2 N PCA, and after a secondcentrifugation the supernatant fluid is completely re-moved. Two samples are used to determine protein(19); dried E. coli protein is used as a standard. Twoother pellets are hydrolyzed in 2 ml of 0.3 N NaOH forI hr at 37 C, after which I ml of cold 1.2 N PCA isadded. After centrifugation, the ultraviolet absorbancyat 260 nm of the supernatant fluid is measured (7, 8);the RNA content of the cells is calculated from thisvalue, using an extinction coefficient of 10.5 mmole-for a mixture of the four 3'-ribonucleoside monophos-phates corresponding to the composition of ribosomalRNA (23). After RNA extraction, the pellet whichcontains deoxyribonucleic acid (DNA) and other cellcomponents is washed once with 2 ml of ice-cold 0.2 NPCA and dissolved in 0.1 ml of 0.06 N NaOH; 0.9 mlof TSM buffer, pH 7.3 (9), and 50 Mg of DNase I(Worthington, I Y crystallized) are then added, and(Worthington, Ix crystallized) are then added, andtation of the proteins with 0.5 ml of 0.2 N PCA (30 min,0 C) and centrifugation, the concentration of oli-godeoxynucleotides is determined spectrophoto-metrically at 260 nm and the DNA content of the cells iscalculated by using an extinction coefficient of 10mmole- 1. Hypochromicity is disregarded since it isless than 5% at the acid pH used (40).

Qualitative estimation of the Km values for variouscarbon sources: the substrate gradient test. A few thou-sand cells are uniformly spread on petri dishes (diame-ter 8 cm), containing 25 ml of minimal agar with thy-mine but without a carbon source. A droplet of a 20%solution or some crystals of the substrate to be testedare then deposited on the surface; during subsequentincubation, a concentration gradient is formed. Cellswith a low Km value for the substrate tested rapidlyproduce colonies at the low substrate concentrations farfrom the "substrate spot," whereas cells with an in-creased Km grow more slowly. The distance from thespot to the furthest visible colonies ("the colony front")is measured at half-day intervals (Fig. 3). A 20-folddifference in the apparent Km is easily registered.

Selection of mutants with an iocreased Km for theuptake of glucose. Cells (109) of a clone of E. coli B/rthy- growing exponentially in AB medium with 0.2%glucose were mutagenized with nitrosoguanidine (4).After washing with minimal medium, the cells weregrown in AB medium with 0.5% glucose. Exponentialgrowth was maintained for 20 generations by appro-priate dilutions with fresh medium to reduce the fre-quency of slowly growing cells and to eliminate auxo-trophs. The desired mutant was expected to grow asrapidly as the wild-type cells at a glucose concentrationof 0.5%. Finally, 109 washed cells were resuspended in

879VOL. 107, 1971

on April 18, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

VON MEYENBURG

250 ml of AB medium with 20 mg of glucose per liter.This concentration is about 100 times the Km value ofthe wild type for growth on glucose; mutants with in-creased Km values should grow slowly, and, therefore,penicillin selection (13) against the wild type should bepossible. A glucose concentration of 10 to 15 mg wasmaintained by appropriate addition of sterile glucosesolution to prevent cessation of growth of the wild-typecells. After 90 min of growth at low glucose concentra-tion (doubling time of entire population, 47 min), peni-cillin [10,000 International Units (IU)/ml] was added.After I hr, the viable count was reduced 7,000-fold,and about 103 surviving cells were plated on MacConkeyagar with 0.07% glucose. The development of the colorof the colonies was checked repeatedly between 10 and15 hr of incubation at 37 C. Wild-type colonies developa.red center after 11 to 12 hr; colonies which turned redlater or developed weaker color, or both, were picked,restreaked, and tested in various ways for growth onglucose.

RESULTS

Growth characteristics of mutants with de-creased affinity for the uptake of glucose. Among24 colonies showing delayed coloring onMacConkey plates with 0.07% glucose, fourwere found to have a considerably increased ap-parent Km value for growth on glucose. In ABmedium with 0.2% glucose, all four grew withthe same doubling time as the wild type (42 min);however, after a 40-fold dilution into AB mediumto reduce the glucose concentration to 50 mg/liter, doubling times of 100, 120, 150, and 180min, respectively, were recorded. The mutant(CP 367) with the highest apparent Km for glu-cose growth is slightly larger and has a broadersize distribution than the wild type when growingat 1.4 to 1.5 doublings/hour in AB medium with0.2% glucose. The three other mutants showedabout a twofold increased cell size. With respectto the uptake of sugars and amino acids, thesefour mutants exhibit the same pleiotropic increaseofKm as is shown below for CP 367.

Figure IA shows the growth of CP 367 at var-ious glucose concentrations. When these culturesapproach the stationary phase, their growth ratesdecrease gradually even at the relatively high ini-tial glucose concentration of 0.1%. Moreover,when a culture growing at 0.1 % glucose (dou-bling time, 43 min) is diluted to give 100, 50, or20 mg of glucose/liter, much lower growth ratesare immediately established. At the same lowglucose concentrations, the parent strain CP 366(Fig. lB) continues growth at its maximum rate(u = 1.5) until the glucose is practically used up.In the mutant CP 367, the same high growth ratecan only be maintained at glucose concentrationshigher than 0.1%.

In the wild type, zt begins to decrease at about

2 mg of glucose/liter (Fig. 2A). At such lowconcentrations, a defined growth rate can only bemeasured in very dilute cultures by following theincrease in cell numbers, e.g., with a Coultercounter. The Km for growth of CP 366 on glu-cose is about 10-6 M; for another E. coli B/r,which is lac-, it was estimated at 2 x 10-6 M.Strains E. coli B and E. coli 15 TAU-bar (9)have Km values below 5 x 10-6, as judged fromOD450 measurements. For strain CP 367, a valueof 0.7 x 10-3 M can be read from the plot of ,uversus glucose concentration S from the recip-rocal plot (Fig. 2B). The linearity between ,u-rand S-1 in the latter agrees with Monod's con-clusion (24) that the relationship between u andS can be formulated in analogy with simple en-zyme kinetics.There is no gross temperature dependence of

the Km value for glucose; the same values arefound at 30 and 37 C (Fig. 2B). Also, the yield,in terms of OD, is the same for mutant and wildtype growing with maximal growth rate: 0.22 to0.25 OD,50 units per 100 mg of glucose per liter.

Pleiotropy of the mutation in CP 367. Thegrowth of CP 367 has been studied with severalsugars and carbon sources. The progression ofthe "'colony front" on minimal plates with dif-ferent substrates is illustrated in Fig. 3A, andactual measurements are shown in Fig. 3B forthree different carbon sources. Delayed appear-ance of mutant colonies (compared to wild type)at a certain distance from the substrate spot isthought to reflect their slower growth at lowconcentrations of the substrate and therefore anincreased Km for growth on that substrate.The following sugars and other simple carbon

sources were tested: D-glucose, lactose, D-fruc-tose, D-galactose, D-mannose, D-mannitol, mal-tose, D-raffinose, L-rhamnose, L-arabinose, D-ribose, glycerol, pyruvate, acetate, and succinate.All of them behaved qualitatively like glucose,and so did the amino acids which can supportgrowth as carbon sources: i.e., aspartic acid,asparagine, glutamine, glutamic acid, trypto-phan, serine, proline, alanine, lysine, leucine,tyrosine, and cysteine. Thus, the Km values for theuptake of many carbon sources, including aminoacids, appear to be increased in strain CP 367.

Quantitative evidence for a decreased affinityfor the uptake of sugars was obtained by deter-mining the growth rates at different concentra-tions and analyzing the data, as illustrated inFig. 2 and 6 for glucose and lactose, respectively.Table I shows that the estimated Km values for allthe sugars tested are from 40 to 1,000 times higherin CP 367 than in the parent strain. For most ofthese substrates, the Km values of the parentalstrain are too low to be determined accurately by

880 J. BACTERIOL.

on April 18, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

VOL. 107, 1971 TRANSPORT MUTANT OF E. COLI

A 1000

I" 500

250

B 100,*@. 5

*50

l00 2000@ 000(0TIME) __-

100 .-

t,- 50 -

20

I4

IXT200 300

T IME(mi)

FIG. 1. Growth of CP 366 and CP 367 at differentglucose concentrations. (A) Cultures of the mutant CP367 in AB medium with 1,000, 500, and 250 mg ofglucose/liter, respectively, inoculated from a culturepregrown with 1,000 mg/liter. The 100, 50, and 20 mgof glucose/liter cultures were obtained by appropriatedilution of portions of the 1,000 mg/liter culture (ar-rows) into prewarmed medium without carbon source.(B) Growth of the wild type CP 366 with 50 and 100mg ofglucose/liter, respectively.

OD measurements, and several of the values inTable I are therefore upper limits.

The characteristics of the amino acid uptakewere obtained from the rates of incorporationinto protein of radioactive amino acids added atdifferent concentrations to exponentially growingglucose cultures (Fig. 4 and 5). The half-satura-tion constants for the uptake of several aminoacids were estimated from such labeling kinetics(Table 2), and 20- to 50-fold increases were reg-istered in agreement with the qualitative resultsobtained with the substrate gradient test. Fromthe reciprocal plots, it appears that the maximalrate of incorporation of leucine or histidine inglucose cultures with a doubling time of 43 minis only about 50% of the theoretical maximaluptake rate (Vmax) found by extrapolation (Fig.4C and 5). The Km for leucine uptake was deter-mined independently by growth rate measure-ments on a leucine-requiring derivative of CP367 at different leucine concentrations. IdenticalKm estimates were obtained with the twomethods (Fig. 5).

Finally, the dependence of growth rate on theexternal concentration of different ions wasstudied. For SOQ42 and PO43 - a definite increase

GLUCOSE CONCENTRATION (mg/')5

GLUCOSE CONCENTRATION 1 (mMM1)

FIG. 2. Relationship between glucose concentration and growth rate for CP 366 and CP 367. (A) Growth ratesin doublings per hour, at 37 C, plotted as a function of the glucose concentrations. The upper scale (0 to 10 mg ofglucose/liter) refers to cultures of the wild-type CP 366 (A) in which the growth rates were determined with aCoulter counter. The lower scale (0 to 4,000 mg of glucose/liter) refers to CP 367 (0) and CP 366 (A) growingat glucose concentrations at which growth was measured in terms of OD4,0. (B) The data from A for growth ofCP 367 at 37 C (0), and additional data for growth at 30 C (A), plotted as p- I (doubling time, g, in minutes)versus (glucose concentration)- 1. At both temperatures, the Km for growth ofCP 367 on glucose is estimated at 7X 10-4 M.

6,lQ

EGo

.200

O ,'00I

.0 so

881

10

A 1.0LUJ

I-

3:0c' 0.50D

on April 18, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

VON MEYENBURG

60

40

20

A

BsQ Q QB

TIME (days)

FIG. 3. Substrate gradient test of the growth of CP366 and CP 367 with different substrates. (A) Sche-matic presentation of the substrate gradient test. S,substrate spot. The small dots indicate colonies which,as incubation is prolonged, become visible at increasingdistance (arrow) from S. (B) Progression of thecolony front (as distance from the substrate spot) withtime of wild-type CP 366 (filled symbols) and CP 367(open symbols) with glucose (0), lactose (0), or pro-line (A) as substrate.

of the apparent Km for growth (or uptake) wasobserved, whereas for the cations K+, NH,+, andMg2+ there was no difference between wild typeand mutant (Table 3).A case of special interest is the growth on lac-

tose. As in the wild type, the Km of the mutantfor growth on lactose is considerably higher thanfor any other sugar (Table 1). With strain CP367, a lactose concentratio-f of 1% is required toreach the growth rate of the wild type; at lowerconcentrations, the growth rate is reduced (Fig.6A), and, as in the case of glucose, the g- Iversus S-1 relationship is approximately linear(Fig. 6B).The high Km value for lactose is not due to

impaired function of the lac operon, resulting inlow levels of j-galactosidase or M-protein (12).Addition of isopropyl-p-D-thiogalactoside (5 x10-4 M) to a culture growing in 0.15% lactosewith a growth rate of 0.5 doublings/hour in-creases the s-galactosidase level about fivefoldbut does not affect the growth rate. Analogousresults were obtained by comparing the growth

on L-arabinose of CP 367 (araC+) with that of aderivative of CP 367 (araCc) constitutive for thearabinose enzymes, the specific permease, andthe binding protein (39).The lactose concentrations at which the growth

rates are greatly reduced are still comparativelyhigh (e.g., 0.15% for a doubling time of 120min). At this low growth rate, a cell density of3 x 108 cells/ml can therefore be reached without,u changing significantly as a result of substrateconsumption.

Revertants. Cultures of strain CP 367 can bemaintained in media with rate-limiting substrateconcentrations only as long as they remain genet-ically stable. Under these conditions, wild-typerevertants would have an obvious selective ad-vantage and would soon outgrow the mutant.To estimate the reversion frequency, cultures

grown up from a single colony in 1% lactosewere used to inoculate 30-ml volumes of minimalmedium with 0.08% lactose. The initial densitywas approximately 107 cells/ml, and the dou-bling time was about 240 min. Usually, ,u beganto increase after about 20 hr, indicating that fastgrowing revertants were taking over. The rever-sion frequency was calculated to be no higherthan 10-8/generation. Thus, cultures with lim-

TABLE 1. Km for growth on carbohydrates of wild-typeCP 366 and mutant CP 367 and uptake of

nonmetabolizable sugarsa

Substrate CP 366 CP 367

D-G1UCOSe .......... 10-6 7 x 10-4Glucose-6-phosphate .2 x 10-6 2 x l0-3D-Galactose6 ....... 10-4 3 x 10-3D-Fructose .........

VOL. 107, 1971 TRANSPORT MUTANT OF E. COLI

TIME (min)

ECL

u

HISTIDINE CONCENTRATIONI( pM 1)

FIG. 4. Histidine uptake by CP 366 and CP 367.Incorporation of 14C-L-histidine (4.5 mCi/mmole) bythe mutant CP 367 (A) and the wild type CP 366 (B),respectively, at 5 (A), I (0), 0.25 (A), 0.1 (0), and0.05 (0) tig of histidine/ml. At the time of histidineaddition, the OD4.0 values of the cultures were 0.175(CP 367) and 0.205 (CP 366), respectively. (C) Graphi-cal estimation of the Km for histidine uptake. The ratesof histidine incorporation (V) are calculated as micro-moles per (gram of protein x minute) from the incor-poration data in A and B and plotted as V- I versus thereciprocal of the histidine concentration. Km for histi-dine uptake: CP 366 (A) < 1.5 x 10- 7 M; CP 367 (0) =3 X 10-6 M. The extrapolated Vmax for histidine uptakeby CP 367 is 6.5 .tmoles per (gram of protein x min-ute).

iting lactose concentrations and with values aslow as 0.2 to 0.25 remain stable for several dou-bling times. For most purposes, a culture may beconsidered to have reached a steady state ofgrowth after three to four mass doublings at con-stant u.To test whether the increased Km value for the

uptake of the sugars, amino acids, and anions isdue to a single mutation, 25 clones of CP 367were grown in 10-ml volumes of AB mediumwith 0.05% lactose (doubling time, about 400min). After 40 hr, these cultures were diluted l/50

into the same medium; 30 hr later, all except twoof the cultures were strongly enriched for rever-tants, as judged from their growth rate in the lowlactose medium. Each "reverted" culture wasstreaked on broth plates, and isolated colonieswere tested on lactose gradient plates. Some ofthe isolated clones appeared to have Km valuesfor growth on lactose as low as the wild type;

LEUCINE CONCENTRATION l(,um 1

FIG. 5. Uptake of leucine by CP 367 (A) andgrowth of CP 367 leu- on leucine (0), plotted as inFig. 4C and 2B, respectively. The Km for leucine uptakeas well as for growth on this amino acid is estimatedat 2.5 x 10- 5 M. Extrapolated Vmax for leucine uptake= 18 umoles per (gram ofprotein x minute).

TABLE 2. Km for the uptake of L-amino acids, uracil,and uridine by wild-type CP 366 and mutant CP 367a

Substrate CP 366 CP 367

Arginine .............. 10-7 4 x 10-6Histidine ...............

VON MEYENBURG

TABLE 3. Km for the uptake (respective growth on) ofions by the wild type CP 366 and the mutant CP 367a

Substrate CP 366 CP 367

P04 -.......

TRANSPORT MUTANT OF E. COLI

port maximal growth rate, whereas in the latterit is the leucine concentration in the cells whichlimits the growth rate. Finally, an increasedcounterflow apparently does not yield the ob-served g versus S relationship, as indicated bythe results obtained with a mutant of E. coli ML308 which appeared to be leaky for galactosides(47).No permease, transport protein, or other

factor has yet been described in which a muta-tional change could be expected to exert thepleiotropic effect on transport seen in strain CP367. The proteins of the phosphotransferasesystem have been shown to be involved in theactive transport of carbohydrates [see review byRoseman (34)]. They are specific either (i) for asingle sugar and under substrate-specific control(enzyme II complex, factor III), or (ii) for amore or less defined group of carbohydrates (en-zyme I, HPr), which does not include glucose-6-phosphate and pentoses (44) for which the Km ismarkedly increased in our mutant (Table 1).Also, enzyme I and HPr lesions have beenmapped at 46 min (45), whereas the mutation inCP 367 maps between strA and metB. Variousproteins, which bind different substrates withhigh specificity, have been isolated from the su-pernatant fluid of osmotically shocked cells (16)and appear to be located in the periplasmic spacebetween the cytoplasmic membrane and the cellwall (15, 27, 32). Dissociation constants of 1O-7to 10-6 M have been reported for binding pro-teins specific for leucine (1, 33), arginine (46),galactose and glucose (1), arabinose (39), SO42-(31), and PO43- (22). No common element seemsto govern the synthesis of this class of proteinssince substrate-specific control of the synthesis ofthe leucine- and arabinose-binding proteins hasbeen reported (33, 39). With arabinose constitu-tive (araCc) derivatives of strains CP 366 and367, we could show that they produce arabinose-binding protein in similar amounts and withcomparable Km values; in contrast, the Km valuefor growth on L-arabinose is 100-fold increasedin the mutant (Table 1).

By definition, transport proteins are situatedon or in the cytoplasmic membrane, i.e., they arestructurally more or less fixed. All transport thustakes place in a highly complex protein-lipidmembrane structure, and any mutation affectingthe cooperativity in the system may have pleio-tropic effects. A mutant has been isolated with arequirement for fatty acid for lactose transport(11), suggesting that the lactose permease, theM-protein of Fox and Kennedy (12), has tocombine with fatty acids to associate with themembrane in the proper active configuration.Similarly, the function of a whole class of spe-

cific binding proteins might be impaired by achange in a structural membrane protein, or by amutation affecting membrane assembly or theinteraction between cell wall and membrane. Infact, it might be impaired by any change in theenvironment in which the binding proteins func-tion.The structure of the cell membrane or wall, or

both, may actually be changed in strain CP 367since these cells are slightly bigger and more het-erogenous in size than wild-type cells. Moreover,ethylenediaminetetraacetic acid (EDTA, 10 mM)causes lysis of the mutant, but not of the wild-type cells. (The other three mutants isolated to-gether with CP 367 show similar pleiotropic in-creases of their Km values; however, they are notEDTA sensitive and the cells are unusually big.)This EDTA sensitivity reverts together with theother characteristics. We first envisaged a changeof the cell envelope (membrane or wall; 29)which would result in loss of the binding pro-teins, but no arabinose-binding activity could bedemonstrated in the supernatant fluid of CP 367araCc cultures. Alternatively, a mutationalchange of the cytoplasmic membrane might in-terfere with the passage of the binding proteinsfrom the cytoplasm where they presumably areproduced into the periplasmic space. However,arabinose-binding activity could be released byosmotic shock (28) from the araCc derivatives ofboth CP 366 and 367 without concomitant lossof ,B-galactosidase, indicating that in both strainsthe binding proteins are located in the peri-plasmic space. [The araCc derivatives were testedby Robert Schleif who also carried out some ofthe arabinose-binding protein determinationswith the equilibrium dialysis assay (39).]By varying the degree of osmotic shock, it was

found that the arabinose-binding protein is easierto release from the mutant than from the wildtype (Fig. 7). This could mean that the bindingproteins are relatively loosely attached to theperiplasmic membrane in the mutant. This inturn could impair their function which presum-ably is to trap substrate molecules and turn themover to the transport chain.The mutation in strain CP 367 seems not to

affect the outer surface of the cell, since themutant can be efficiently synchronized by theHelmstetter-Cummings membrane elution tech-nique (14), which requires that the cells stick tocertain membrane filters; E. coli B/r, from whichCP 367 is derived, sticks particularly well.

Since the functional analysis shows that theKm of the mutant for the transport of a widespectrum of nutrients is increased without re-vealing the exact nature of the mutationalchange, I designate its genetic locus kmt.

885VOL. 107, 1971

on April 18, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

value characteristic of the initial substrate con-4 - centration. At low cell densities, a lactose con-

centration of 0.13% generates a gs value of 0.6,and, below an OD,,0 value of 0.550, M > 0.5.The corresponding figures for a lactose concen-

3 tration of 0.07% are: M = 0.3 at low cell densi-ties, and ,u > 0.24 below an GD,,0 value of0.220. As indicated by the Km values in Table 2,

2 / / the corresponding substrate concentration and2 / upper limits are 10 times lower when glucose isused as carbon source. Higher cell densities maybe reached, but this requires carefully calculated

1 . and timed addition of substrate to keep i withinthe desired interval. Furthermore, experimentsinvolving prolonged growth at low rates are onlypossible if the level of revertants is kept low byusing as inoculum a freshly isolated clone.0510 15 20 The RNA/DNA and RNA/protein ratios for

1. SUCROSE strain CP 367 are shown in Fig. 8; they representbatch cultures in which the growth rate was de-

G. 7. Release of L-arabinose-binding activity fined by the lactose concentration. The results!wild-type CP 366 and mutant CP 367 by different agree quite well with the data of Rosset et al.ees of osmotic shock. Arabinose constitutive (35) and Forchhammer and Lindahl (10) on theCC) derivatives of CP 366 and CP 367 were grown composition of E. coli as a function of theIB minimal medium with 0.4% glycerol to an gowth of En t as depends on the510 of2.8, at which 100 ml of each culture was cen- growth rate when the latter depends on the ac-ged. The cells were washed, and five portions of tual substrate used, and not on its concentration.strain were osmotically shocked as described by The overall composition of the cell thus varieseif (39), except that the sucrose concentration was similarly with the growth rate, whether this is5, 5, 10, and 20%, respectively, during the incuba- determined by the rate at which a given substrate

tion at 22 C in 1 ml of 0.03 M Tris-hydrochloride (pH7.3)-10-4 M EDTA; the cells were then centrifuged andresuspended in 0.6 ml of ice-cold 10'4 M MgCl,. Wevary the degree of osmotic shock in proportion to thesucrose concentration (abscissa). Volumes (0.3 ml) ofthe shock fluids of CP 366 araCe (A) and CP 367araCc (-) were dialyzed for 90 min at 4 C against 3 x10- I M 14C-I -arabinose (3 mCi/mmole); 14C counts perminute were determined in 0.1-ml volumes of the dia-lyzed samples and dialysis buffer (dashed line). Countsin the shock fluids exceeding the counts in the bufferreflect the amount of arabinose-binding protein re-leased by osmotic shock.

Application of the mutant to growth studies.The increase of the Km for the uptake of manysubstrates makes strain CP 367 extremely usefulfor the study of several problems related to bac-terial growth. Steady-state growth up to rela-tively high cell densities can be obtained atgrowth rates as low as 0.2 doublings/hour bylimiting the concentration of a single substrate.It is obvious that, under such conditions, thesubstrate concentration and, thus, the growthrate, must decrease continuously as the culturegrows. From the relationship between substrateconcentration and At (determined at low cell den-sities) and the yield in cell mass per milligram ofsubstrate, we can calculate the cell density belowwhich g will remain within, e.g., 10% of the

10

S0

44zz

cca

0.5 1.0 1.5

0.4 _

C

0.2

6

GROWTH RATE Y(hr-1)

FIG. 8. Cell composition of the mutant CP 367 asfunction of the growth rate. RNAIDNA (A) andRNA Iprotein (0) ratios were determined on cell sam-ples harvested from batch cultures of CP 367 growingat different lactose concentrations. Wild-type CP 366(A, 0) growing on 0.2% lactose has the same cellcomposition as CP 367 with I% lactose.

x

E

0

Fjfromdegr4(ara(in AOD4,trifugeachSchk0, 2..

6 0~~~~a a

7 0-0,

0/

886 VON MEYENBURG J. BACTERIOL.

2

on April 18, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

TRANSPORT MUTANT OF E. COLI

is taken up or by the efficiency with which agiven substrate can support catabolism. Thissuggests that the growth rate is controlled at alevel at which it does not matter whether one oranother set of enzymes serves to maintain thenecessary flow of carbon and energy; a modelwhich satisfies this condition has recently beenproposed (20).The general correspondence between growth

rate and cell composition is particularly relevantfor experiments involving rate changes. Withstrain CP 367, a simple change of the substrateconcentration (by dilution or addition) will pro-duce shifts which do not involve derepression orrepression of the synthesis of new substrate spe-cific enzyme systems. A shift-down of this kindis illustrated in Fig. 1. The dilution with minimalmedium, without substrate, replaces the other-wise necessary filtration, washing, and resuspen-sion of the culture. This simple procedure alsohas the advantage that the new growth condi-tion (i.e., the new substrate concentration) isimmediately established. In chemostat cultures, ashift performed by reducing or increasing themedium flow rate introduces a prolonged tran-sient state (26), because in this case the substrateconcentration is dependent on the growth re-sponse of the cell population.

Strain CP 367 also lends itself to studies ofphosphate- and sulfate-limited growth rates; withsuitable auxotrophs, amino acid-limited growthrates can be maintained in batch culture at usefulcell densities. In this laboratory, the mutant hasbeen successfully applied to studies of cell size asa function of lactose- or leucine-limited growthrates, yielding results similar to those obtainedby Schaechter et al. (37) for Salmonella typhi-murium. Finally, the timing of DNA synthesis inthe division cycle at doubling times up to 250min has been investigated by Jesper Zeuther (un-published data) with the membrane elution tech-nique of Helmstetter and Cummings (14).

ACKNOWLEDGMENTS

I am very much indebted to 0. MaalOe for initiating sugges-tions and for helpful criticism throughout the course of thisstudy, to R. Schleif for discussions and introducing me to the"secrets of the arabinose-binding protein," and to Lissie Dahl-gren for excellent technical assistance.

This investigation was supported by grants from the Euro-pean Molecular Biology Organization (EMBO).

LITERATURE CITED

1. Anraku, Y. 1968. Transport of sugars and amino acids inbacteria. 1. Purification and specificity of the galactose-and leucine-binding proteins. J. Biol. Chem. 243:3116-3122.

2. Beck, C., and K. von Meyenburg. 1968. Enzyme patternand aerobic growth of Saccharomyces cerevisiae undervarious degrees of glucose limitation. J. Bacteriol. 96:479-485.

887

3. Boyer, H. 1966. Conjugation in E. coli. J. Bacteriol. 91:1767- 1772.

4. Cerda-Olmedo, E., P. C. Hanawalt, and N. Guerola. 1968.Mutagenesis of the replication point by nitrosoguanidin:map and pattern of replication of the E. coli chromo-some. J. Mol. Biol. 33:705-719.

5. Clark, J., and 0. Maal#e. 1967. DNA replication and thedivision cycle in E. coli. J. Mol. Biol. 23:99-1 12.

6. Ecker, R. E., and M. Schaechter. 1963. Ribosome contentand the rate of growth of Salmonella typhimurium.Biochim. Biophys. Acta 76:275-279.

7. Fleck, A., and D. Begg. 1965. The estimation of ribonu-cleic acids using ultraviolet absorption measurements.Biochim. Biophys. Acta 108:333-339.

8. Fleck, A., and H. N. Munro. 1962. The precision of ultra-violet absorption measurements in the Schmidt-Tannhauser procedure for nucleic acid estimation.Biochim. Biophys. Acta 55:571-583.

9. Forchhammer, J., and N. 0. Kjeldgaard. 1967. Decay ofmessenger RNA in vivo in a mutant of E. coli 15. J.Mol. Biol. 24:459-470.

10. Forchhammer, J., and L. Lindahl. 1971. Growth rate ofpolypeptide chains as function of the cell growth rate ina mutant of E. coli 15. J. Mol. Biol. 55:563-568.

11. Fox, C. F. 1969. A lipid requirement for induction of lac-tose transport in E. coli. Proc. Nat. Acad. Sci. U.S.A.63:850-855.

12. Fox, C. F., and E. P. Kennedy. 1965. Specific labeling andpartial purification of the M-protein, a component of the,8-galactoside transport system of E. coli. Proc. Nat.Acad. Sci. U.S.A. 54:891-899.

13. Gorini, L., and H. Kaufmann. 1960. Selecting bacterialmutants by the penicillin method. Science 131:604-605.

14. Helmstetter, C., E., and D. J. Cummings. 1964. An im-proved method for the selection of bacterial cells at divi-sion. Biochim. Biophys. Acta 82:608-610.

15. Heppel, L. A. 1967. Selective release of enzymes from bac-teria. Science 156:1451-1455.

16. Heppel, L. A. 1969. The effect of osmotic shock on releaseof bacterial proteins and on active transport. J. Gen.Phys. 54:95S- I 13S.

17. Koch, A. L., and C. S. Deppe. 1971. In vivo assay of pro-tein synthesizing capacity of E. coli from slowly growingchemostat cultures. J. Mol. Biol. 55:549-562.

18. Lindahl, L., and J. Forchhammer. 1969. Evidence for re-duced breakdown of messenger RNA during blockedtranscription or translation in E. coli. J. Mol. Biol. 43:593-606.

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

20. Maal0e, 0. 1969. An analysis of bacterial growth. De-velop. Biol. Suppl. 3:33-58.

21. Maalge, O., and N. 0. Kjeldgaard. 1966. Control of mac-romolecular synthesis. W. A. Benjamin, New York.

22. Medveczky, N., and H. Rosenberg. 1969. The binding andrelease of phosphate by a protein isolated from E. coli.Biochim. Biophys. Acta 192:369-371.

23. Midgley, J. E. M. 1962. The nucleotide base compositionof RNA from several microbial species. Biochim. Bio-phys. Acta 61:513-525.

24. Monod, J. 1942. Recherches sur la croissance bacteriennes.Herrmann & Cie., Paris.

25. Monod, J. 1950. La technique de culture continue. Theorieet applications. Ann. Inst. Pasteur 79:390-410.

26. Mor, J. R. 1969. Growth of Saccharomyces cerevisiaeunder transient conditions. Proc. 4th Int. Symp. Contin-uous Cultivation of Microorganisms. Academia, Prague,p. 297-307.

27. Nakane, P. K., G. E. Nichoalds, and D. L. Oxender. 1968.Cellular localization of leucine-binding protein from E.coli. Science 161:182-183.

VOL. 107, 1971

on April 18, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

VON MEYENBURG

28. Neu, H. C., and L. A. Heppel. 1965. The release of en-zymes from E. coli by osmotic shock and during theformation of spheroplasts. J. Biol. Chem. 240:3685-3692.

29. Normark, S. 1969. Mutation in E. coli K 12 mediatingspherelike envelopes and changed tolerance to ultravioletirradiation and some antibiotics. J. Bacteriol. 98:1274-1277.

30. Novick, A., and L. Szilard. 1950. Experiments with thechemostat on spontaneous mutations of bacteria. Proc.Nat. Acad. Sci. U.S.A. 36:708-719.

31. Pardee, A. B. 1966. Purification and properties of a sul-fate-binding protein from S. typhimurium. J. Biol.Chem. 241:5886-5892.

32. Pardee, A. B., and K. Watanabe. 1968. Location of sul-fate-binding protein in S. typhimurium. J. Bacteriol. 96:1049-1054.

33. Penrose, W. R., G. E. Nichoalds, J. R. Piperno, and D. L.Oxender. 1968. Purification and properties of a leucine-binding protein from E. coli. J. Biol. Chem. 243:59215928.

34. Roseman, S. 1969. The transport of carbohydrates by abacterial phosphotransferase system. J. Gen. Phys. 54:138S-184S.

35. Rosset, R., J. Julien, and R. Monier. 1966. Ribonucleicacid composition of bacteria as function of growth rate.J. Mol. Biol. 18:308-320.

36. Rotman, B., A. K. Ganesan, and R. Guzman. 1968. Trans-port systems for galactose and galactosides in E. coli. 11.Substrate and inducer specificities. J. Mol. Biol. 36:247-260.

37. Schaechter, M., 0. Maaloe, and N. 0. Kjeldgaard. 1958.Dependency on medium and temperature of cell size and

chemical composition during balanced growth of S. ty-phimurium. J. Gen. Microbiol. 19:592 606.

38. Schleif, R. 1967. Control of production of ribosomal pro-tein. J. Mol. Biol. 27:41-55.

39. Schleif, R. 1969. An I-arabinose binding protein and arabi-nose permeation in E. coli. J. Mol. Biol. 46:185-196.

40. Shugar, D. 1960. Photochemistry of nucleic acids and theirconstituents, p. 55. In E. Chargaff and J. N. Davidson(ed.), The nucleic acids, vol. 3. Academic Press Inc.,New York.

41. Sykes, J., and T. W. Young. 1968. Studies on the ribo-somes and RNA of Aerobacter aerogenes in carbon-limited continuous culture. Biochim. Biophys. Acta 169:103-116.

42. Taylor, A. L. 1970. Current linkage map of E. coli. Bac-tefiol. Rev. 34:155-175.

43. von Meyenburg, K. 1969. Katabolit-Repression und derSprossungszyklus von Saccharomyces cerevisiae. Viertel-jahresschr. Naturforsch. Ges. Zurich 114:113-222.

44. Wang, R. J., and M. L. Morse. 1968. Carbohydrate accu-mulation and metabolism in E. coli. 1. Description ofpleiotropic mutants. J. Mol. Biol. 32:59-66.

45. Wang, R. J., H. G. Morse, and M. L. Morse. 1969. Car-bohydrate accumulation and metabolism in E. coli:close linkage and chromosomal location of ctr muta-tions. J. Bacteriol. 98:605-610.

46. Wilson, 0. H., and J. Holden. 1969. Stimulation of argi-nine transport in osmotically shocked E. coli W cells bypurified arginine-binding protein fractions. J. Biol.Chem. 244:2743-2749.

47. Wong, P. T. S., E. R. Kashket, and T. H. Wilson. 1970.Energy coupling in the lactose transport system of E.coli. Proc. Nat. Acad. Sci. U.S.A. 65:63-69.

888 J. BACTERIOL.

on April 18, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/