JOURNAL OF BACrERIOLOGY, Oct. 1973, p. 397-409Copyright 0 1973 American Society for Microbiology

Vol. 116, No. 1Printed in U.S.A.

Membrane Synthesis in Synchronous Cultures ofBacillus subtilis 168

MICHAEL G. SARGENT

National Institute for Medical Research, Mill Hill, London, NW7 IAA, England

Received for publication 13 June 1973

Synthesis of bacterial membranes has been investigated in Bacillus subtilis byexamining incorporation of amino acids and glycerol into the protein and lipid ofmembranes of synchronous cultures. A simple reproducible fractionation schemedivides cellular proteins into three classes (i) truly cytoplasmic, (ii) looselymembrane bound, released by chelating agents, and (iii) tightly membranebound. These comprise approximately 75, 10, and 15%, respectively, of cellularproteins in this organism. Incorporation of radioactivity into these fractions,using steady-state and pulse labeling has been followed during the cell cycle.Cytoplasmic proteins and the loosely membrane-bound proteins are labeled at anexponential rate throughout the cell cycle. The membrane fraction is labeleddiscontinuously in the cell cycle, with periods of rapid synthesis over the latterpart of the cycle and a period with no net synthesis during the early part of thecycle. Pulse labeling indicates that synthesis of membrane occurs at a linear ratethat doubles at a fixed time in each cycle, which coincides with the period of zeronet synthesis. Rates of membrane synthesis measured by pulse labeling duringthe period of rapid membrane synthesis are significantly less than indicated bysteady-state labeling. These discrepancies are consistent with the hypothesisthat during the cell cycle certain proteins are added to the membrane from thecytoplasm and that during the period of zero net synthesis there is an efflux ofproteins from the membrane. Evidence in favor of this has been presented. Theactivity of succinic dehydrogenase (a representative of class c) varies in a

step-wise manner with periods of rapid increase, approximately coincident withbursts of membrane protein synthesis, alternating with periods without anyincrease in activity. The activities of malate dehydrogenase (class a) and reducednicotinamide adenine dinucleotide dehydrogenase (class b) increased throughoutthe cell cycle. Phospholipid synthesis is continuous throughout the cell cycle.

At a morphological level, synthesis of cellmembrane involves (i) continuous extension ofperipheral membrane between division, (ii) for-mation of mesosome and cross-wall membrane,and (iii) separation of deoxyribonucleic acid(DNA)-membrane attachment site into twowith apparent extension of the membrane be-tween sites. To relate morphology to the bio-chemical basis of membrane assembly, compo-nent insertion must be considered in relation tocell age (6, 11, 16, 17, 32) and cellular location(1, 3, 5, 9, 12, 19, 24, 26, 28, 38, 43, 47). Moststudies of envelope synthesis have been con-fined to the question of location of sites ofsynthesis. Membrane synthesis may occur atmany dispersed foci or in an equatorial band.The latter possibility, with its corrollary that

old envelope components are conserved, is thebasis of the replicon theory of genome segrega-tion (13). Incorporation of each membrane com-ponent should be considered separately sinceseveral studies suggest that incorporation ofphospholipid and protein is not mutally de-pendent (14, 25).With one exception (28), data for phospholip-

ids (9, 19, 26, 43, 47) favor randomly dispersedincorporation. The possibility that mesosomalphospholipid may be the precursor of plasmamembrane has been eliminated (6, 26, 28).Tumover, the thermal mobility of phospholip-ids, and their chemical heterogeneity, may tendto obscure a specific focus of incorporation.

Autoradiographic or immunofluorescenttechniques are probably inherently impossible

397

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

with membrane proteins, but indirect experi-ments in which segregation of certain induciblecharacters are studied indicate conservation ofmembrane protein and therefore localized inser-tion (1). Autoradiographic studies of incorpo-rated 3H-a-aminolaevulinic acid segregationsuggest dispersed incorporation of membranecytochrome proteins but with conservation ofquite small membrane subunits (9). Cell wallsynthesis must to some extent reflect mem-brane synthesis, and immunofluorescent tech-niques indicate that the wall is synthesized inan equatorial band in many strains and that,once synthesized, it is conserved through sev-eral generations (5, 12). Chai and Lark (4) claimthat a complete round of DNA synthesis isconserved and segregated, with cell wall synthe-sized one generation later in Lactobacillusacidophilus. A specific site of autolysis in Esch-erichia coli has been regarded as evidence oflocalized wall synthesis (8, 41). Autoradio-graphic studies of cell wall synthesis favor bothmodels (3, 19, 24).Very few biochemical studies of envelope

synthesis in relation to cell age have been made.Discontinuous synthesis of mucopeptide (11)and phospholipid (6) have been claimed. Re-cently, Ohki (32) has demonstrated discontinu-ous increases in cytochrome bl and glycero-phosphate transport with a simultaneous in-crease in turnover of phosphatidyl glycerol.Kubitschek et al. (16, 17) consider that uptakesites for all nutrients double in activity at celldivision.This communication describes membrane

synthesis in relation to the cell cycle of Bacillussubtilis.

MATERIALS AND METHODSBacteria, growth conditions, and synchronous cells

were described (M. G. Sargent, J. Bacteriol., to bepublished).

Preparation of cell fractions. Samples of bacterialculture (1-2 ml) were transferred to ice-chilled centri-fuge tubes containing chloramphenicol (CAM) (100tg/ml). Carrier to dilute isotopes was added where

required. Cells were sedimented and suspended inprotoplast medium (PM) containing sucrose (0.5 M),magnesium sulfate (0.02 M, pH 6.5), malate buffer(0.02 M), and lysozyme (100 jug/ml) and were incu-bated at 35 C for 15 min. Under these conditionsprotoplast formation is always complete. Protoplastswere sedimented at 15,000 rpm for 15 min. Themesosomal vesicles (FI) were sedimented at 100,000 xg for 2 h. Since this fraction contains only 1% of cellprotein, this step was omitted in certain experiments,in which case, protoplasts were prepared in a smallvolume (0.2 ml) and were lysed by dilution withtris(hydroxymethyl)aminomethane (Tris)-mag-

nesium (0.05 M Tris, pH 7.4, containing 0.02 Mmagnesium sulfate). After removal of the mesosomes,protoplasts were suspendA in 0.2 ml of PM and werelysed in Tris-Mg in a similar way. Samples were leftfor 1 h in ice to give maximal lysis, and the mem-branes were sedimented and washed in Tris-Mg.Supernatant fluids were pooled (FII). Sediments werewashed twice in citrate-phosphate (trisodium citrate,0.01 M, and pH 7.4 phosphate buffer, 0.1 M). Super-natant fluids were pooled (FIII), and sediments (FIV)were suspended either in citrate-phosphate (for en-zyme assays) or sodium hydroxide (0.1 M) for estima-tion of radioactivity.

Radioactive methods: protein labeling. "C- and3H-tryptophan or -histidine were used to label proteinas shown in text. In the presence of 2 and 5 Ag oftryptophan and histidine, respectively, per ml incor-poration is proportional to optical density (OD) up to1.2. Samples for counting (1 ml, in duplicate) werevigorously mixed with 10% trichloroacetic acid (2 ml)after addition of carrier amino acid (200 ug/ml) andbovine serum albumin (100 Ag/ml), and were heatedat 80 C for 30 min, collected and dried on WhatmanGF/A filters, and counted in toluene scintillation fluid(2 ml). Protein synthesis in samples was stopped withCAM (100 /g/ml) and chilling.

Pulse labeling. Samples (1 ml) of a synchronousculture were transferred to prewarmed, boilingtubes containing 1 gCi of 3H-tryptophan. Sampleswere incubated at 35 C with shaking for 11 min, afterwhich the pulse was terminated by addition of 1 ml ofan ice-cold solution containing CAM (200 Ag/ml),sodium azide (0.002 M), and tryptophan (200 Ag/ml).Cells were sedimented and suspended in citrate-phos-phate containing lysozyme (200 Ag/ml, 0.25 ml).When lysis was complete, samples were stirred vigor-ously with a Vortex mixer and were diluted with 2 mlof citrate-phosphate.The membranes (FIV) were then sedimented and

washed once more together with carrier membranes(50 jig of protein). Sediment and supernatant fluid(membranes and cytoplasm, respectively) were proc-essed as above.Membrane turnover in synchronous cultures.

Bacteria (500 ml) were grown in the presence of 0.2,uCi of 3H-tryptophan per ml and 3 jlg of nonradioac-tive tryptophan per ml until cells reached an OD (540nm) of 2.0. Cells were then diluted into 500 ml of freshmedium. Synchronous cells were prepared immedi-ately and were resuspended in fresh medium.

Lipid labeling. Glycerol-2-3H was used to labelphospholipid (in the presence of 50 ,g of glycerol perml) as described by Daniels (6). To stop incorpora-tion, 1 ml of an ice-cold solution of GAC (glycerol, 2mg/ml; sodium azide, 0.002 M; CAM, 200,ug/ml) wasadded to an equal volume of sample. Cells weresedimented and resuspended in GAC containing lyso-zyme (100 jig/ml) and then incubated for 30 min at35 C. Two volumes of ice-cold trichloroacetic acid(10%) were added, and after 30 min the precipitatedphospholipids and protein were collected on GF/Afilters. Filters were washed with distilled water andthen dried. To express tritium counts fully, the filterswere immersed in 2 ml of ethanol, incubated at 100 C

398 SARGENT

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MEMBRANE SYNTHESIS IN B. SUBTILIS

until dry, and counted with toluene scintillation fluid. pared as described above, before and after citrateTrichloroacetic acid-precipitable radioactivity pre- phosphate treatment, were mixed with an equalpared by this method represented total phospholipid volume of 5% glutaraldehyde and left at room temper-glycerol and was not significantly contaminated by ature for 1 h. The membranes were then sedimentedglycerol teichoic acids. After lysozyme treatment, and washed five times with distilled water. Samplesradioactivity that was not precipitated with ice-cold were mixed with equal volumes of 2% phosphotung-trichloroacetic acid (10%) was not extracted into stic acid (pH 7.0), 3% sodium silicotungstate (pH 7.0),chloroform-methanol (2:1, vol/vol). Trichloroacetic or 4% ammonium molybdate (pH 7.0), stained for 5acid-precipitated radioactivity was almost completely min, and examined at 60 kV on a Phillips EM 300extracted into chloroform-methanol. The amount of electron microscope.radioactivity in such a trichloroacetic acid precipitate Gel electrophoresis. After dialysis against distilledwas almost equal to the amount extracted into water and freeze drying, protein samples for electro-chloroform-methanol without prior trichloroacetic phoresis were heated for 5 min at 90 C in 1% sodiumacid precipitation. Greater precision in measuring dodecyl sulfate and 1% mercaptoethanol. Polyacryla-phospholipid synthesis in synchronous cells was ob- mide gel electrophoresis (46) was carried out by usingtained using the trichloroacetic acid precipitation 140 lsg of protein on 7.5% gels. Proteins in the gel wereprocedure rather than by extraction into chloroform- stained with amido-black.methanol.Enzyme assays: preparation of lysates. Samples RESULTS

(1 ml) were placed in ice-cold centrifuge tubes con-taining CAM (100 Mg/ml). Cells were sedimented and Preparation and characterization of cellsuspended in citrate-phosphate containing CAM (100 fractions. In analyzing the synthesis of the,ug/ml) and lysozyme (100 tsg/ml). Control experi- bacterial cell membrane, it was desirable toments indicated that when CAM (100 ug/il) was divide cellular proteins into three classes: (i)added to growing cultures at 35 C there was no further cytoplasmic soluble proteins that are releasedincrease in succinic dehydrogenase activity after 2 by osmotic lysis of protoplasts, (ii) typical

Succinic dehydrogenase (EC 1.3.99.1). Reactin insoluble membrane proteins, and (iii) proteinsmixtures in 1-ml cuvettes contained 0.2 ml of sodium that are loosely or adventitiously attached tosuccinate (0.5 M, adjusted to pH 7.4), 0.05 ml of the membrane. Published data suggest thatdichlorophenol-indophenol (1 mM), euzyme. prepara- chelating agents would release ribosomes andtion, and citrate-phosphate to give a volume-of 0.95 certain proteins from membranes (29, 30). Theml. Assays were started by using 0.05 ml of phepazine effect of repeated washings in Tris-Mg andmethosulfate (3 mg/ml-freshly prepared) and wmeree citrate-phosphate are illustrated in Fig. 1. Pro-read against a blank lacking phenazine methsul fte. toplasts, prepared as above from bacteria la-A unit of enzyme activity decreases the absobanceof beled with '4C-tryptophan and 3H-glyceroldichlorophenol-indophenol by 0.1 at 600 nm id 1 nIin w sat 22 C. The reaction rate was unaffected-by cyanide.

Malate dehydrogenase (EC 1.1#I.37). Reaction medium to achieve lysis. After 2 h of incubationmixtures in 1-ml cuvettes, contained 0.1 ml of at 0 C under these conditions to give maximalnicotinamide adenine dinucleotide (NAD) (1 mM), lysis, the particulate fraction was sedimented at0.05 ml of dichlorophenol-indophenol (1 mM), en- 15,000 rpm for 20 min (supernatant fluid fromzyme preparation, and citrate-phosphate buffer to this step = wash 1). The sediment was thengive a volume of 0.8 ml. Assays were started with 0.2 washed three more times. The percentages ofml of sodium malate (0.5 M, adjusted to pH 7.4), total cellular protein, phospholipid, and malatereading against a blank lacking substrate. A unit Of dehydrogenase in the membrane are shown.enzyme activity decreases absorbance of dichloro- Although almost all malate dehydrogenase isphenol-indophenol by 0.1 at 600 nm in 1 mi at 22 C. released at the first step, the amount of mem-Cyanide and phenazine methosulfate had no signifi- rane te irst doe not reachm-cant effect on the reaction rate. brane protein in Tris-Mg does not reach aNADH-dehydrogenase. Reaction mixtures con- constant level until the sediments have been

tained 0.2 ml of reduced NAD (NADH) (10 mM), 0.1 washed four times. Membranes in citrate-phos-ml of potassium cyanide (10 mM), enzyme prepara- phate reach constant protein composition aftertion, and citrate-phosphate buffer to give a volume of two washes. Phospholipid was released from0.95 ml. Assays were started with 0.05 ml of dichloro- membrane by successive washes, although withphenol-indophenol (1 mM), reading against a blank citrate-phosphate the rate of release was threelacking substrate. A unit of enzyme activity decreases times that with Tris-Mg. Continuous washing ofabsorbance of dichlorophenol-indophenol by 0.1 at membrane with Tris-Mg apparently ultimately600 nm in 1 min at 22 C. measep eth are loosely ultomteChemical assays. Ribonucleic acid (RNA) and releases proteins that are loosely bound to theprotein were determined colorimetrically by the membrane.methods of San-Lin and Schjede (40) and Lowry et al. Table 1 illustrates certain analytical features(20), respectively. of the fractionation that support this point of

Electron microscopy. Samples of membrane, pre- view. Thus, malate dehydrogenase is found in

399VOL. 116, 1973

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

FII (Tris-Mg-soluble extract) almost exclusive-ly, and succinic dehydrogenase is almost exclu-sively in FIV (citrate-phosphate insoluble). FIII(extracted from Tris-Mg membranes with cit-rate-phosphate) has the highest specific activ-ity of NADH-dehydrogenase, although it ispredominantly an FII enzyme. Citrate-phos-phate treatment evidently removes most of theRNA from the membranes and presumablyfrom ribosomal proteins. When examined by

FIG. 1. Effect of washing on membrane fractioncomposition. A lysate, prepared as in text, labeledwith '4C-tryptophan and 3H-glycerol was washed as

shown with either Tris-Mg or citrate-phosphate.Points indicate percentage of component remaining in15,000 rpm sediment after each wash. Symbols: O, *,malate dehydrogenase in sediment in either Tris-Mg-or citrate-phosphate-washed membrane; 0, *, 14C-protein in Tris-Mg or citrate-phosphate-washedmembrane; A, A, 3H-glycerol in Tris-Mg- or citrate-phosphate-washed membrane.

polyacrylamide gel electrophoresis in sodiumdodecyl sulfate, distinct differences are seen

between FIII and FIV, although certain bandsare clearly present in both fractions. The stain-ing patterns of gels prepared using the same

amount of protein (140,g) from fractions IIIand IV are illustrated in Fig. 2. Bands 2, 3, 6, 9,and 11, which are major components of FIV,are present in reduced amounts or are com-

pletely absent from FIII. Bands 4, 5, and 7 arepresent apparently in similar amounts in bothfractions. Fraction III is probably slightly en-

riched in band 8. Differences in minor bandsare also evident. The visible bands in FIII donot appear to contribute sufficient protein tocompensate for the bands that are absent. Thediscrepancy may be accounted for by a largenumber of proteins in low concentrations de-rived from ribosomes and perhaps contaminat-ing cytoplasmic protein.The appearance under the electron micro-

scope of magnesium-stabilized and citrate-treated membranes, negatively stained withphosphotungstic acid, is illustrated in Fig. 3.Similar results were obtained with silicotung-state and ammonium molybdate. The proto-plast ghost under both conditions is completewith very little indication of fragmentation ofits basic structure. The membrane surface,however, appears to become more granular aftercitrate treatment. After two washings in citrate-phosphate, the membrane forms one majorband in density gradients.

Steady-state incorporation of protein labelinto cell fractions. The relationship between"4C-tryptophan incorporation into each cellfraction and the OD of the culture is illustratedin Fig. 4. In each case, the radioactivity permilliliter of culture increased in proportion tothe increase in OD, with no lags apparent.

Incorporation of '4C-tryptophan into synchro-nous cells is illustrated in Fig. 5. Radioactivityincorporated into FII and FIII increased contin-uously and in proportion to OD, indicating an

exponential rate of increase. However, countsincorporated into FIV, plotted against time,

TABLE 1. Characteristics of cell fractions

Malate dehydrogenase Succinic dehydrogenase NADH dehydrogenaseFraction Protein RNA

(%of total) (% of total) % Sp act p%S act S%p act(U/mg) (U/mg) 0 (U/mg)

II 73 87 99 1.67 0 0 68 3.03III 10 11 1 0.01 5.5 0.28 24 8.48IV 16 2 0 0 94.5 3.02 9.3 1.96

400 SARGENT

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MEMBRANE SYNTHESIS IN B. SUBTILIS

o-4

,kt-

2345

6

78910I I

.'

lop

12 -- _-

13

FE FIZ

FIG. 2. Gel electrophoresis of citrate-soluble and-insoluble fractions of the membrane.

showed periods of rapid incorporation (0-50 and90-120 min) alternating with periods without orwith reduced synthesis (55-90 and 125-145min) (Fig. 5).With '4C-histidine as a protein label, similar

results were obtained, but the discontinuity ofmembrane labeling was obscured if the bacte-ria, from which synchronous cells were selected,

were not grown in histidine for several genera-tions.Rates of membrane synthesis by pulse

labeling. Although steady-state measurementsof total membrane synthesis (Fig. 5) indicate aperiod in the cell cycle in which there is no netsynthesis of membrane, substantial rates of3H-tryptophan incorporation into FIV were ob-served in short pulses. The relationship betweenthe pattern of steady-state and pulse labelingwas investigated by using the same synchronousculture. This was divided into two parts (i)which was steady-state labeled by growing cellsin the presence of 3H-tryptophan (0.5,Ci/ml)and (ii) which was pulse labeled for 11 minevery 5 min (1 ,Ci/ml). The radioactivity incor-porated into membrane (FIV) and cytoplasmic(FII and FIII) protein was determined and isshown in terms of nanograms of 3H-tryptophan.The specific activity of 3H-tryptophan was cal-culated as described above. The rate of cyto-plasmic protein synthesis obtained by pulselabeling indicated an exponential increase inthe rate of synthesis comparable with the ratecalculated from steady-state labeling, althoughsubstantial differences were observed betweenrates of membrane synthesis obtained by pulseand steady-state labeling (Fig. 6). The rate ofmembrane synthesis calculated from steady-state labeling, assuming a linear rate of mem-brane synthesis (shown as a cross-hatched his-togram in Fig. 6), was approximately constantduring the intervals 0 to 65 min and 100 to 130min, with no net synthesis in the intervals 65 to100 and 130 to 160 min. Pulse labeling alsoindicates an approximately constant rate ofmembrane synthesis during the intervals 0 to 70and 100 to 160 min, although at an apparentlylower rate of synthesis than indicated bysteady-state labeling. In sharp contrast to thesteady-state labeling data, the rate of membranesynthesis actually increases during the intervals70 to 100 and 160 to 190 min. The linear ratesare approximately 30, 52, and 110 to 130 ng permin per ml, and these are, relative to the secondplateau, in the ratio 1.16:2:4.6 which approxi-mates to doublings in the rate. The pulse data,therefore, suggest that membrane is synthesisedat a constant linear rate that doubles at aparticular time in the cell cycle (cell age-35min approximately).The time course of membrane accumulation,

assuming a rate of membrane synthesis equiva-lent to the pulse rate of membrane synthesis, isshown in Fig. 6 superimposed on the data forsteady-state labeling. At the end of the period of

VOL. 116, 1973 401

I

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

402 SARGENT J. BACTERIOL.

A BFIG. 3. Electron micrographs of membrane fractions. (A) Magnesium stabilized; (B) citrate treated.

Glutaraldehyde-fixed preparations, negatively stained with phosphotungstic acid. Bar 0.5 gm.

w

2 X IdU-o 6

~~2g~~~~~c~~11 2

_ . H * | ~~~~~r_ a 2 ..

0~~~~~~~~~~~~~~~CO4QPICAL 0:6 DENSITY 54Onm 0'8 x r

FIG. 4. Incorporation of '4C-trypohn nocll )fractions relative to optical density. Symbols: *,cytoplasm (P11); 0, citrate-insoluble fraction (FfV0; -30@, citrate-soluble fraction (FIII).

ITIME 2 HOURSzero net synthesis, the amount of membranepresent in steady-state labeled cells (24 and 56 -yptophan incorpora-ng of membrane tryptophan at 90 and 160 mmn, tion into cell fractions of a synchronous culture.

*~~~ ~ ~~~~~~~~Smos 0,0c an,..4Ctr

respectively) is very similar to the amount that yptophan incorporatedwouldhaveaccuulate on he bsis o therateinto trichloroacetic acid-precipitable protein in FII,woulsnhaesi andccuuated on thle lbasisgof2th rate FIlI, and FIV, respectively, per ml of culture; -,ofsntheisidicaed b pule laelin (27andoptical density (540 nm); 0, Coulter count per milli-

62 ng of membrane tryptophan). These observa- liter of culture; U, Coulter count corrected for num-tions are consistent with the following hypothe- bers of units with septa.

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MEMBRANE SYNTHESIS IN B. SUBTILIS

sis. During the period of net membrane synthe-sis, newly synthesised protein and relatively oldproteins from the cytoplasm are incorporatedinto the cell membrane. At the end of the periodof net synthesis, an efflux of protein occurswhich is numerically equivalent to the amountof "old" protein added to the membrane fromthe cytoplasm.Membrane turnover in synchronous cells.

The hypothesis, suggested above to explain thediscrepancy between pulse and steady-statelabeled data can be tested by pulse-chase stud-ies of membrane synthesis. Two predictionsfollow from the hypothesis.

(i) If membranes are labeled during the pe-riod of rapid synthesis and are then chased withnonradioactive tryptophan at the end of theperiod of rapid synthesis, there should be anefflux of radioactivity during the period of zeronet synthesis. The amount of protein involved isequivalent to the difference between the steady-state level of membrane radioactivity and thevalue calculated on the basis of linear synthesisat the rate obtained by pulse labeling. In thecase of the discontinuities shown in Fig. 6, theamount of protein in the efflux is 29 and 27% ofthe amount present at 65 and 130 min, respec-tively.

(ii) If cells are pulse labeled and then chasedduring the period of synthesis, the radioactivityof the membrane fraction should increase dur-ing the chase and then decline when the periodof no net synthesis starts.These postulates were tested as follows. A

synchronous culture was labeled with 1 ,uCi of3H-tryptophan per ml in the presence of 3 ,ug ofcarrier tryptophan per ml for either 50 or 70 minfrom the time of selection. To terminate thepulse, samples of culture were collected onmembrane filters (Millipore Corp.), washed, andsuspended in fresh media containing 3 ,ug ofnonradioactive tryptophan. Levels of mem-brane and cytoplasmic radioactivity were moni-tored during synchronous growth after eithertreatment. The level of radioactivity in mem-brane and cytoplasm in both cases is illustratedin Fig. 7. In the case of cells labeled for 70 minafter selection (Fig. 7A), membrane-bound ra-dioactivity falls sharply over the first 10 min,during which approximately 30% of the radioac-tivity is lost, and then decreases more slowly.Cytoplasmic protein turnover occurs at about12% per generation. In cells labeled for 50 min(Fig. 7B), the amount of membrane-boundradioactivity rises by about 18% over 5 min,stabilizes for 30 to 40 min and then starts to fall.Cytoplasmic protein turnover is approximately6 to 8%.

FIG. 6. Comparison of rates of membrane synthesisobtained from steady-state and pulse labeling. Sym-bols: 0, rate of membrane synthesis obtained by pulselabeling (log scale) (cross-hatched area, rate of mem-brane synthesis obtained from steadv-state labeling);0, time course of steadv-state membrane synthesis(arithmetic scale);. , calculated time course ofmembrane synthesis based on rate of membranesynthesis obtained from pulse experiments; -,Coulter count per milliliter; -, Coulter count permilliliter corrected for numbers of units with septa;- -, optical density at 540 nm.

These experiments suggest that relativelynewly synthesised proteins turn over during theperiod of no net synthesis, but do not considerchanges in older cell proteins. To study these,cells were labeled with 3H-tryptophan prior toselection of small cells (see Materials andMethods) and were washed and resuspended infresh medium containing nonradioactive tryp-tophan. The radioactivity of cytoplasmic pro-tein decreases exponentially at the rate of 8%per generation. However, the radioactivity ofmembrane protein decreases in a step-wisefashion at an average rate of 12% per generation(Fig. 8). The peak rate of membrane turnover isobserved midway through the cell cycle (cell age30 min) during the period of zero net synthesis.

403VOL. 116, 1973

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

404

100

80

*->100

a 800

av2 100,

800a)N

4_1000

0l~80-

SARGENT

0A0 00 0 100°C

*. B

'.00c 00300

0 0D\0. 0

oaco

o , s

0 30 TIME 6bMINS gbFIG. 7. Turnover of membrane label during non-

radioactive chase of synchronous cells. Syvnchronouscells were labeled for 70 min (A and B) or 50 min(C and D) with 3H-tryptophan from time of selection,after which they were collected on membrane filters(Millipore Corp.), washed, and suspended in growthmedia. Symbols: 0, cytoplasmic protein, 0. rnem-brane protein.

Enzyme synthesis in synchronous cells.Activities of succinic, malate, and NADH dehy-drogenase per milliliter of synchronous cultureare illustrated in Fig. 9 and 10, together withdata for changes in cell number. NADH dehy-drogenase and malate dehydrogenase activitiesincrease throughout the cell cycle (Fig. 9).Assays for these enzymes are not sufficientlyaccurate to distinguish between continuous,linear synthesis and exponential synthesis. Suc-cinic dehydrogenase activity (Fig. 10) increasesin a step-wise fashion with maximal rates ofincrease at 15 to an and 110 to 125 min. The firstperiod of synthesis occurs marginally earlierthan septum formation and the second occursslightly after cell separation. Periods of oxygendeprival (up to 15 min) cause no discontinuityin synthesis of these enzymes in unselectedexponentially growing cultures.

Total phospholipid synthesis. Phospholip-ids were labeled with glycerol-2-3H. Incorpora-tion was proportional to the OD of cultureswhile in exponential growth in the presence of50 ,ug of nonradioactive glycerol per ml. Daniels(6) has argued that this isotope labels onlyphospholipids because any reactions involvingthe 2 position result in loss of the 3H-label to the

J. BACTERIOL.

cell water pool. However, in B. subtilis 168labeled with glycerol-2- 3H, radioactivity isfound in the glycerol teichoic acid and phospho-lipid. These can be separated analytically byprecipitation of phospholipid in a lysozymelysate with cold trichloroacetic acid (see Mate-rials and Methods). Protein is not labeled underthese conditions.The time course of total glycerol incorpora-

tion into the phospholipids of synchronous cellsis illustrated in Fig. 11, together with cellnumber data. Incorporation is continuous and ifplotted against OD is seen to be exponential.

DISCUSSIONThe significant role played by plasma mem-

brane and mesosome in bacterial cell divisionindicated by electron microscopy has not beenadequately described in molecular terms. Theavailability of methods for analyzing cell struc-ture (35) and obtaining synchronous cultures

HOURSFIG. 8. Turnover of membrane protein labeled

prior to selection of synchronous cells. Syvmbols: 0,membrane protein (FIV); 0, cytoplasmic protein(pooled FII and FIII); -, Coulter units per milliliter;-, Coulter units corrected for number of cells withsepta; - -, optical density (540 nm).

IIV

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MEMBRANE SYNTHESIS IN B. SUBTILIS

38); maximal yields being favored by low/6 magnesium levels. The level used in these" 1 EJ experiments (20 mM) is the lowest commensu-/ rate with minimal leakage of cytoplasmic pro-

n.- teins and at which release of vesicles seems to be32 /,vf08- 16 almost complete morphologically (35). The mes-

*/p osomal bag in this case is almost certainly inEn the FIV fraction.04 Assembly of the membrane in relation to the- z cell cycle has been studied by following incorpo-zz ration of tryptophan or histidine into the pro-

:D84 0 -4( :teins of cell fractions in synchronous cells.Eu ' o-< wL Tryptophan is of special value for such experi-Nz:DD ments since it is absent from the flagella of B.4o o subtilis (21), and feasible methods of separating0 60 120 180 flagella from membrane are not available..9.MUsing steady-state labeling, fractions FIl andFty Flll were synthesised exponentially, but FIVin synchronous cultures. Malate dehydrogenase (0) . . . '

and NADH-dehydrogenase (x 10-2) (0) activities per was synthesised in an apparently discontinuousmilliliter of culture. Symbols: -, optical density; , fashion with periods of incorporation alternat-Coulter count per milliliter; *, Coulter count cor-rected for numbers of units with septa.

(M. G. Sargent, submitted for publication) hasmade this more feasible with bacilli. A repro-ducible procedure has been used that separatesthe cell proteins into three fractions with mini-mal cross-contamination. Fraction II is cyto-plasmic and fraction IV is the residue of the 0plasma membrane after extraction with chelat-ing agents. Under the electron microscope this Eappears as a protoplast ghost, and there is no cevidence of gross fragmentation, although the 7 9surface of the membrane is altered in texture. v/ oThis is consistent with the view that citrate- o 0 20solubilized proteins (FIII) may be located on themembrane surface. Enzymes, including NADHdehydrogenase, that are loosely attached to theplasma membrane have been demonstrated byothers (7, 29, 30).Owing to the serious difficulties in obtaining 0D10210

quantitative yields of mesosomal material, a I /clear-cut pattern of mesosome synthesis has notbeen obtained. The cause of this probably lies in /rthe heterogeneity of membrane structures,within the mesosome. Rogers (36) has empha- E D

sized that the mesosome is composed of an ,cintracytoplasmic membrane (that is continuous Go0 0415with the cell membrane) and vesicles and tu-bules enclosed within it, which may be attachedto the membrane to varying extents. Procedures 1 HOURS 2 3for obtaining mesosomes yield only the vesicu- FIG. 10. Succinic dehydrogenase activity in syn-lar component, and no convincing isolation of chronous cultures. Symbols: M, succinic dehydrogen-the mesosomal "bag" has been described. The ase activity per milliliter; -, optical density; 0,yield of mesosomal vesicles and tubules de- Coulter count per milliliter; *, Coulter count cor-pends on the level of magnesium used (35, 37, rected for numbers of units with septa.

405VOL. 116, 1973

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

FIG. 11. Phospholipid synthesis in synchronouscells. Symbols: U, 3H-glycerol incorporated into tri-chloroa"cetic acid-precipitable fraction of lysozymelysate per milliliter of culture; -, optical density; 0,

Coulter count per milliliter; 0, Coulter count cor-rected for numbers of units with septa.

ing with periods without net incorporation. Incontrast, pulse-labeling experiments indicatethat membrane is synthesized continuously at alinear rate, which increases sharply, at a partic-ular cell age and remains constant until thesame time in the next cycle. Two kinds ofdiscrepancy are evident. When rates of mem-brane synthesis, obtained from pulse labeling or

calculated from steady-state labeling data, arecompared. Firstly, during the period of zero netsynthesis, the measured rate of membrane syn-thesis does not decrease but remains constantand almost doubles in rate towards the end ofthe period. Secondly, during the period ofsynthesis, pulse measurements of the rate ofmembrane synthesis are consistently lower thanthe rate of synthesis calculated from steady-state measurements. Rates of cytoplasmic pro-tein synthesis obtained by the two methodswere virtually identical, suggesting that it isunlikely that the lower rate of membrane syn-thesis obtained by pulse labeling could be atechnical artifact.To account for the first discrepancy, an efflux

of protein-bound radioactivity must be postu-

lated, amounting to approximately 30% of themembrane synthesized in the period of rapidsynthesis. This can be demonstrated formallyby monitoring the decline in membrane-boundradioactivity in the presence of nonradioactivetryptophan in cells labeled prior to the discon-tinuity in membrane synthesis. Thus mem-branes labeled for 70 min after selection ofsynchronous cells show a considerable amountof turnover which is numerically equivalent tothe postulated efflux. Turnover of old mem-brane labeled prior to selection of cells was alsoexamined. In this case periodic bursts of turn-over were observed amounting to approximately12% per generation, coincident with the periodof zero net synthesis. Thus, the postulatedefflux of radioactivity from the membrane isderived principally from relatively newly syn-thesised proteins, although there is a contribu-tion from old membrane. Nishi and Kogoma(31) have demonstrated periodic bursts of pro-tein turnover (10% per generation) coincidentwith septation in E. coli.An explanation of the second discrepancy

(between the steady-state rate of membranesynthesis and the rate of pulse labeling) couldbe that the steady-state rate of membranesynthesis comprises two contributions: (i) pro-tein that is newly synthesised and (ii) olderprotein added to the membrane, presumablyfrom the cytoplasm. Since pulse labeling wouldestimate only the former, the higher rate ofsynthesis obtained in steady-state-labeled cellscould include a contribution from the cyto-plasm. On this view, an upward curve in thetime course of membrane labeling should beexpected and was observed in most steady-statelabeling experiments during the period of syn-thesis, although the line of best fit is uncertain.This hypothesis is also supported by the obser-vation that, if cells are labeled for 50 min duringthe period of rapid synthesis and are thenchased with nonradioactive tryptophan, there isan 18% increase in membrane-bound radioac-tive protein during the first 5 min of the chase.This observation suggests that proteins synthe-sized in the cytoplasm may be transferred to themembrane fraction in the manner suggestedabove. The significance of the two contributionsto the net rate of membrane synthesis cannot beassessed at present. However, the amount ofprotein added to the membrane originating fromthe cytoplasm is numerically equivalent to theefflux of membrane protein that occurs duringthe period of no net synthesis, and also the pro-teins involved in the efflux are labeled princi-pally during the cycle immediately prior to thediscontinuity in net synthesis and not in a pre-

406 SARGENT

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MEMBRANE SYNTHESIS IN B. SUBTILIS

vious cycle. The efflux of radioactivity may in-volve transfer of intact proteins to the cytoplasmor alternatively protein degradation to trichloro-acetic "acid-soluble products. An unequivocaldistinction between the possibilities is not pos-sible as protein degradation clearly occurs underthe circumstances of these experiments. Pine(34) has demonstrated a population of rapidly-turning-over proteins within the cell envelopefraction of E. coli.The principle enzymatic functions of the

bacterial membrane, nutrient transport, andenergy generation are difficult to study in syn-

chronous cultures since large amounts of mate-rial are required. Furthermore, the significanceof measurements of these activities in terms ofamounts of specific proteins is difficult to as-

sess. However, a preliminary approach to thisproblem has been made in this communicationby studying succinic dehydrogenase activity, an

example of an extremely insoluble enzyme.

Malate and NADH dehydrogenases have alsobeen studied since they are representatives ofthe cytoplasmic and insoluble fractions, respec-

tively, and are also respiratory enzymes. Theactivity of succinic dehydrogenase increasesdiscontinuously, with increases at the time ofcell division. Such a change in enzyme activitycan reflect either alterations in the specificactivity of individual molecules or in theamount of enzyme protein. Rigorous discrimi-nation between these alternatives is not possibleat present, although inhibition of protein syn-

thesis by CAM stops the increase in enzyme

activity (in contrast to the adenosine triphos-phatase of Micrococcus lysodeikticus [45], sug-

gesting that no activation of the protein occurs

subsequent to synthesis. In contrast to succinicdehydrogenase, the respiratory components,NADH-dehydrogenase and malate dehydrogen-ase, which are not in the citrate-insoluble frac-tion are synthesised continuously.The activities of stable enzymes in the bacte-

rial cell cycle follow two major patterns (27), (i)continuous linear synthesis, doubling in rate atthe time of duplication of its structural gene

(10, 33), (ii) discontinuous synthesis (as shownby succinic dehydrogenase, above), so thatenzyme activity increases for only a brief periodin the cell cycle (18, 22, 23). "Step" enzymes ofthis kind seem to appear in a sequence corre-

sponding to the gene order, though not coinci-dent with the time of gene duplication and can

appear in the absence of DNA replication. Thephenomenon is not well understood but hasbeen explained in terms of oscillatory repression(22) or linear transcription (42). Recently, thesuggestion has been made that certain step

enzyme phenomena (18) are not observed insynchronous cells obtained by selectionmethods and that the steps may be caused bythe method of synchronization (2). Step en-

zymes in germinating spores are thought toreflect a developmental process that is con-

trolled by linear transcription (42). The step insuccinic dehydrogenase synthesis occurs justbefore cell separation and, therefore, is added inthe course of a linear membrane growth period.Step-wise increases in certain membrane activi-ties have been observed by others. Ohki (32) hasshown that cytochrome bl, an extremely insolu-ble membrane protein in E. coli, increases in a

step-wise fashion in synchronized cells. Bymeasuring rates of uptake of a variety of nutri-ents, Kubitschek (16, 17) and Ohki (32) claim tohave shown that uptake sites (presumably per-

mease molecules) double in number at specifictimes in the cell cycle of E. coli.These observations give a preliminary outline

of membrane assembly in relation to the cellcycle. Five possible processes that may beinvolved in the insertion of protein into themembrane are indicated in Table 2.

Total phospholipid synthesis measured as

glycerol incorporation is apparently continuousin the cell cycle of B. subtilis, in agreement withOhki's (32) observations on E. coli. In contrast,with B. megaterium synchronized by amino acidstarvation, Daniels (6) observed a relativelysmall acceleration in the rate of phospholipidsynthesis at the time of septum formation.Measurements of total lipid synthesis will prob-ably mask specific changes in individual molec-ular species, and in fact Ohki (32) has shownthat phosphatidyl glycerol turnover is restrictedto a short period in the cell cycle.

In any study with synchronous cells, the

TABLE 2. Processes involved in incorporation ofproteins into membranes

Process Cell agea

1. Linear incorporation of newly synthesizedprotein (pulse labeled) with sharp changesin rate approximately 50 min before cellseparation ............................... 3 min

2. Addition of proteins from the cytoplasm tothe citrate insoluble fraction ...... ....... 35-85 min

3. Efflux of proteins from the citrate insolublefraction ................................ 0-35 min

4. Addition of citrate-extractable protein (in-cluding ribosomal protein) ...... ......... Throughout

cycle5. Discontinuous appearance of succinic de-

hydrogenase ............................ 75-10 min

aCells in which daughter cells have just separated are

regarded as at cell age 0. Septum formation occurs 20 minbefore separation (cell age 65 min). Doubling time 85 min.

407VOL. 116, 1973

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

408 SARGENT

possibility exists that phenomena observed arecreated by the synchronization procedure. Al-though the method used here is both rapid andgentle, B. subtilis 168 nevertheless shows asmall reduction in growth rate (up to 105%) afterselection. This organism is extremely delicatewhen grown in minimal media, and reductionsin growth rate can be induced by dilution intolarge volumes of prewarmed media. Publishedestimates of protein turnover in exponentiallygrowing cultures of bacilli have been quite low(2-7%) per generation (44). However, we typi-cally have found higher and rather variablerates of turnover in B. subtilis 168 both inexponential and synchronous cultures. The pos-sibility must be considered that filtration forselection of synchronous cells may induce sometumover. A strain of B. subtilis has now beenisolated in which there is no reduction in massgrowth rate during selection of synchronouscells (unpublished results).

ACKNOWLEDGMENTSAdvice and encouragement from H. J. Rogers, and the

technical assistance of M. F. Bennett are gratefully acknowl-edged. The electron-micrographs were prepared by P. B.Wyrick.

LITERATURE CITED

1. Autissier, F., and A. Kepes. 1971. Segregration of mem-brane markers during cell division in Escherichia coli.II. Segregation of lac permease and mel permeasestudied with a penicillin technique. Biochim. Biophys.Acta 249:611-615.

2. Bellino, F. L. 1973. Continuous synthesis of partiallyderepressed aspartate transcarbamylase during thedivision cycle of Escherichia coli. J. Mol. Biol.74:223-238.

3. Briles, E. B., and A. Tomasz. 1970. Radioautographicevidence for equatorial wall growth in a gram positivebacterium. J. Cell. Biol. 47:786-790.

4. Chai, N., and K. G. Lark. 1967. Segregation of deoxyribo-nucleic acid in bacteria. Association of the segregatingunit with the cell envelope. J. Bacteriol. 94:415-421.

5. Cole, R. M. 1965. Symposium on the fine structure andreplication of bacteria and their parts. III. Bacterial cellwall replication followed by immunofluorescence. Bac-teriol. Rev. 29:326-344.

6. Daniels, M. J. 1969. Lipid synthesis in relation to the cellcycle of Bacillus megaterium KM and Escherichia coli.Biochem. J. 115:697-701.

7. De Siervo, A. J., and M. R. J. Salton. 1971. Biosynthesisof cardiolipin in the membranes of Micrococcuslvsodeikticus. Biochim. Biophys. Acta 239:280-292.

8. Donachie, W. D., and K. J. Begg. 1970. Growth of thebacterial cell. Nature (London) 227:1220-1224.

9. Green, E. W., and M. Schaecter. 1972. The mode ofsegregation of the bacterial cell membrane. Proc. Nat.Acad. Sci. U.S.A. 69:2312-2316.

10. Helmstetter, C. E. 1968. Origin and sequence of chromo-some replication in Escherichia coli B/r. J. Bacteriol.95:1634-1641.

11. Hoffmann, B., W. Messer, and U. Schwarz. 1972. Regula-tion of polar cap formation in the life cycle of Esche-richia coli. J. Supramol. Struct. 1:29-37.

J. BACTERIOL.

12. Hughes, R. C., and E. Stokes. 1971. Cell wall growth inBacillus licheniformis followed by immunofluorescencewith mucopeptide-specific antiserum. J. Bacteriol.106:694-696.

13. Jacob, F., S. Brenner, and F. Cuzin. 1963. On thereplication in bacteria. Cold Spring Harbor Symp.Quant. Biol. 28:329-348.

14. Kahane, I., and S. Razin. 1969. Synthesis and turnover ofmembrane protein and lipid in Mycoplasma laidlawii.Biochim. Biophys. Acta 183:79-89.

15. Kennett, R. H., and N. Sueoka. 1971. Gene expressionduring outgrowth of Bacillus subtilis spores. J. Mol.Biol. 60:31-44.

16. Kubitschek, H. E. 1970. Evidence for the generality oflinear cell growth. J. Theqr. Biol. 28:15-29.

17. Kubitschek, H. E., M. L. Freedman, and S. Silver. 1971.Potassium uptake in synchronous and synchronisedcultures of Escherichia coli. Biophys. J. 11:787-797.

18. Kuempel, P. L., M. Masters, and A. B. Pardee. 1965.Bursts of enzyme synthesis in the bacterial duplicationcycle. Biochem. Biophys. Res. Commun. 18:858-867.

19. Lin, E. C. C., Y. Hirota, and F. Jacob. 1971. On theprocess of cellular divisions in Escherichia coli. J.Bacteriol. 108:375-385.

20. 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.

21. Martinez, R. J., D. M. Brown, and A. N. Glazer. 1967.The formation of bacterial flagella. III. Characterisationof the flagella of Bacillus subtilis and Spirillum serpens.J. Mol. Biol. 28:45-51.

22. Masters, M., and W. D. Donachie. 1966. Repression andthe control of cyclic enzymic synthesis in Bacillussutbtilis. Nature (London) 209:476-479.

23. Masters, M., and A. B. Pardee. 1965. Sequence of enzymesynthesis and gene replication during the cell cycle ofBacillus subtilis. Proc. Nat. Acad. Sci. U.S.A.54:64-70.

24. Mauck,, J., L. Chan, and L. Glazer. 1972. Mode of cellwall growth of Bacillus megaterium. J. Bacteriol.109:373-378.

25. Mindich, L. 1970. Membrane synthesis in Bacillussubtilis. II. Integration of membrane proteins in theabsence of lipid synthesis. J. Mol. Biol. 49:433-439.

26. Mindich, L., and S. Dales. 1972. Membrane synthesis inBacillus subtilis. Ill. Morphological localisation of thesites of membrane synthesis. J. Cell Biol. 55:32-41.

27. Mitchison, J. M. 1971. The biology of the cell cycle.Cambridge University Press, Cambridge, England.

28. Morrison, D. C., and H. J. Morowitz. 1970. Studies onmembrane synthesis in Bacillus megaterium KM. J.Mol. Biol. 49:441-459.

29. Munoz, E., M. S. Nachbos, M. T. Shor, and M. R. J.Salton. 1968. ATPase of Micrococcus lysodeikticus.Selective release and relationship to membrane struc-ture. Biochem. Biophys. Res. Commun. 32:539-546.

30. Nachbar, M. S., and M. R. J. Salton. 1970. Characteris-tics of a lipid-rich NADH dehydrogenase containingparticulate fraction obtained from MicrococcusIvsodeikticus membranes. Biochim. Biophys. Acta223:309-320.

31. Nishi, A., and T. Kogoma. 1965. Protein turnover in thecell cycle of Escherichia coli. J. Bacteriol. 90:884-894.

32. Ohki, M. 1972. Correlation between metabolism of phos-phatidylglycerol and membrane synthesis in Esche-richia coli. J. Mol. Biol. 68:249-264.

33. Pato, M. L., and D. A. Glaser. 1968. The origin anddirection of replication of the chromosome of Esche-richia coli B/r. Proc. Nat. Acad. Sci. U.S.A.60:1268-1274.

34. Pine, M. J. 1965. Heterogeneity of protein turnover inEscherichia coli. Biochim. Biophys. Acta 104:439-456.

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MEMBRANE SYNTHESIS IN B. SUBTILIS

35. Reaveley, D. A., and H. J. Rogers. 1969. Some enzymicactivities and chemical properties of the mesosomesand cytoplasnic membranes of Bacillus licheniformis6346. Biochem. J. 113:67-79.

36. Rogers, H. J. 1970. Bacterial growth and the cell enve-

lope. Bacteriol Rev. 34:194-214.37. Rogers, H. J., D. A. Reaveley, and I. D. J. Burdett. 1967.

The membrane systems of Bacillus licheniformis, p.

303-313. In H. Peeters (ed.), Proceedings of the 15thColloq. Protides Biological Fluids. Elsevier PublishingCompany, Amsterdam.

38. Ryter, A. 1971. Etude de la croissance de la membranechez Bacillus subtilis au moyen de la distribution desflagelles. Ann. Inst. Pasteur 121, no. 3.

39. Ryter, A., C. Frehel, and B. Ferrandes. 1967. Comparte-ment des mesosomes lors de l'attaque de Bacillussubtilis por le lysozyme en milieu hyper- et hypoto-nique. Co. R. Acad. Sci. 265:1259-1262.

40. San-Lin, R. I., and 0. A. Schjeide. 1969. Micro-estima-tion of RNA by the cupric ion catalysed orcinolreaction. Anal. Biochem. 27:473-483.

41. Schwarz, U., A. Asnus, and H. Frank. 1969. Autolyticenzymes and cell division of Escherichia coli. J. Mol.Biol. 41:419-429.

42. Steinberg, W., and H. 0. Halvorson. 1969. Timing ofenzyme synthesis during outgrowth of spores of Bacil-lus cereus. J. Bacteriol. 95:469-478.

43. Tsukayoshi, N., P. Fielding, and C. F. Fox. 1971. Mem-brane assembly in Escherichia coli. I. Segregation ofpreformed and newly formed membrane into daughtercells. Biochem. Biophys. Res. Commun. 44:497-502.

44. Urba, R. 1959. Protein breakdown in Bacillus cereus.

Biochem. J. 71:513-518.45. Vambutas, V. K., and M. R. J. Salton. 1970. Differential

inhibitory effect of chloramphenicol on synthesis ofATPase and cytoplasmic enzymes of Micrococcuslvsodeikticus. Biochim. Biophys. Acta 203:94-103.

46. Weber, K., and M. Osborn. 1969. The reliability ofmolecular weight determinations by dodecylsulphatepolyacrylamide gel electrophoresis. J. Biol. Chem.244:4406-4412.

47. Wilson, G., and C. F. Fox. 1971. Membrane assembly inEscherichia coli. II. Segregation of preformed andnewly formed membrane proteins into cells and mini-cells. Biochem. Biophys. Res. Commun. 44:503-511.

VOL. 116, 1973 409

on April 9, 2020 by guest

http://jb.asm.org/

Dow

nloaded from