Vol. 144, No. 3JOURNAL oF BACTERIOLOGY, Dec. 1980, P. 865-8680021-9193/80/12-0865/04$02.00/0

Effect of Environmental pH on Chain Length of Lactobacillusbulgaricus

S. K. RHEE AND M. Y. PACK*Department ofBiological Science and Engineering, Korea Advanced Institute of Science, Cheongyangni,

Seoul, Korea

Culture medium pH was found to affect strongly the chain length of Lactoba-cillus bulgaricus NLS-4 cells. The organism was cultured continuously in glucose-limited complex medium of different pH's with constant agitation at 250 rpmunder anaerobic headspace. The cell chains increased their lengths with anincrease in pH and yielded clumps of folded filaments at pH above 8.0. Involve-ment of an autolytic enzyme(s) in the separation of L. bulgaricus cells wasconfirmed, and the poor synthesis of this enzyme(s) under alkaline cultureconditions could explain the pH-related filamentous growth of this organism.

It has been frequently observed that certainbacteria grow as long forms with a tendency toelongate into filaments. Fan (5, 6) isolated a cellwall-bound enzynme(s), autolysin(s), from Bacil-lus subtilis and confirmed that the enzyme wasresponsible for unlinking cells from long chains.Involvement of enzymes in the autolysis of cellwall components, thus preventing filamentousgrowth, has been studied by many other workers(3, 8, 12-14); it may be suspected that any cul-ture conditions which have some influence onthe synthesis or activity of these enzymes mayconsequently extend their effects to the mode ofcell arrangement for these organisms. However,no systematic study of the relationship betweenculture conditions and the filamentous growthof bacteria has been reported yet. The purposeof our experiments was to determine whetherthe pH of culture medium could regulate thechain lengths in Lactobacillus bulgaricus NLS-4.

MATERIALS AND METHODSOrganism. We used L. bulgaricus strain NLS-4 in

all experiments. This strain, in the lyophilized state,was obtained from W. E. Sandine (Department ofMicrobiology, Oregon State University, Corvallis).The organism was reactivated and maintained in El-liker lactic broth (4). Inoculum for the continuousculture experiments was prepared by overnight culti-vation in the same medium used for the experiments.Growth medium. A semisynthetic medium con-

taining 1% tryptone, 0.5% yeast extract, 0.05% (vol/vol) Tween 80, and 0.5% (vol/vol) each of stock solu-tions A and B was developed by modifying the mediaoriginally formulated by Koditschek et al. (7) and deVries et al. (2) and used throughout the study. SolutionA was composed of 10% K2HPO4 and 10% KH2PO4and solution B consisted of 4% MgSO4.7H20, 0.2%MnSO4-4H20, 2% NaCl, and 0.2% FeSO4.7H20. Glu-cose was sterilized separately and added to give a final

concentration of 10 mM. All ingredients except inor-ganic salts were obtained from Difco Laboratories,Detroit, Mich.Continuous culture. A New Brunswick BioFlo

C30 chemostat (New Brunswick Scientific Co., NewBrunswick, N.J.) with a working volume of 360 ml wasused. The culture was initiated batchwise by inocula-tion with 5 ml of an overnight culture. When thegrowth reached the end of the log phase, a continuousfeeding of fresh medium to the culture was started andadjusted to give a constant dilution rate of 0.23 h-'throughout the experiment. The pH level in the cul-ture solution was controlled automatically with theaddition of 1 N KOH and 1 N H2SO4 (pH 4.0 to 8.5)by an automatic pH controller (model pH-40, NewBrunswick Scientific Co.). Each pH shift required theworking volume of the flow of medium to turn over atleast three times to reach a steady state at whichsamplings were made for analyses. Other conditionsfor the continuous culture were fixed as follows: agi-tation at 250 rpm, temperature at 38° C, and dissolvedoxygen concentrations at zero. The anaerobic condi-tion was maintained by passing a stream of purenitrogen through the headspace of the culture vessel.To assure the anaerobic condition, the dissolved oxy-gen level in the culture solution was periodicallychecked with a dissolved oxygen analyzer (model DO-50, New Brunswick Scientific Co.). Foaming was pre-vented by automatic addition of 1:100 dilution of sili-cone oil by an automatic foam controller (model AFP-101, New Brunswick Scientific Co.).Growth measurement. The course of growth and

the establishment of steady state conditions were de-termined by frequent samplings of the culture andmeasurements of their turbidities. Optical density at660 nm was measured by a Spectronic 20 colorimenter(Bausch & Lomb, Inc., Rochester, N.Y.) calibrated tothe dry cell weight in grams per liter. To determinethe dry cell weight, cells were harvested by centrifu-gation at 27,000 x g for 2 min and washed three timeswith cold distilled water. Washed cells were dried to aconstant weight at 1050C.Measurement of chain length and photomi-

crography. The lengths of 50 individual chains were865

866 RHEE AND PACK

measured by a calibrated micrometer and the rangeand standard deviation of chain lengths were calcu-lated. Photomicrographs were taken with a phase-contrast microscope (American Optical Corp., Buffalo,N.Y.).

Preparation of cell wall extracts and substratefilaments. The cells were harvested when continuouscultures reached steady states and then treated soni-cally by a Sonic 300 dismembrator (Artek SystemsCo., Farmingdale, N.Y.) at 7 kc for 10 min; disruptedcell walls were collected and extracted by proceduresdescribed by Fan (5). Long-chained cells which hadbeen grown under pH 7.5 were harvested and sus-pended in 0.1 M Tris-HCl buffer (pH 7.5). The cellfilaments in this preparation had an average length of100,m, were stable at 5° C, and were free from shorten-ing by mechanical agitation; thus, they were used as asubstrate for the enzymatic dechaining reaction.

RESULTSEffect of medium pH on chain length.

Figure 1 shows a typical batch-culture patternofL. bulgaricus NLS-4 in a glucose-limited com-plex medium. The parallel reverse sigmoidalcurves of pH in the culture solution and thechain length of the cells suggest some closerelationships between the two variables. Thiswas confirmed by the continuous culture data(Table 1). When the pH of the culture mediumwas fixed at 4.5, most of the cells in the suspen-sion remained single or paired, showing an av-erage chain length of about 8 ,um. The chainlength increased with the increase of pH due tothe addition of cells to the chain. When the pHreached 8.0, the cell chains became folded intoclumps resembling thread balls, and theirlengths could not be determined. Under thesehigh pH levels, the growth of the organism wasvery poor, and the cells were washed out when

6.0

0.3Z 1a0

.. . , , .700 1"to~~~~1-101z0

6.0 A02 8_

ZO

TIME hr.)FIG. 1. Changes ofpH, chain length, and dry cell

weight in batch culture. L. bulgaricus NLS-4 wasgrown under conditions of 38° C, 250 rpm agitation,and anaerobic headspace. Symbols: A, pH; *, chainlength; , dry cell weight. Cells having an averagechain length of 1OO,um were inoculated.

TABLE 1. Effect of culturalpH on chain length andcell number per chain of L. bulgaricus NLS-4 in a

steady-state continuous culturepH of culture No. of cells per chaina Chain lengthmedium (a) (I,b

4.0 6.0 ± 1.1c 304.5 1.5 ± 0.6 7.55.5 4.0 ± 1.0 206.5 7.5 ± 1.6 37.57.5 20.0 ± 2.6 1008.0 Clumps _d8.5 Clumps

a Average cell number per chain of 50 chains.ba x 5 lam (average length of unit cell was 5 t,m).C Standard deviation of the distribution.d Impossible to estimate.

the pH exceeded 8.5, regardless of a reduction ofthe dilution rate from 0.23 to 0.03 h-. The elon-gation of filaments in accord with the rise ofenvironmental pH is more clearly shown in Fig.2. No matter how long the filament extended,the unit cell length of 4 to 5 um remained un-changed. When the long-chained cells were re-turned to the lower pH medium, the chain lengthwas reduced again, and no sign of genetic changecould be detected.Involvement of enzymes in chain-length

determination. Previous reports (3, 5, 6, 8, 13,14) show that some enzymes attached to the cellwalls of certain bacteria are responsible for theseparation of cells. This possibility in L. bulgar-icus was examined. The cells grown at pH 7.5with an average chain length of 100 ,um weretreated with fresh cell wall extracts preparedfrom short-chained cells grown at pH 5.5. Thereactions proceeded in buffer solutions ofvariouspH's. As shown in Fig. 3, the fresh extractsdechained the cells effectively, with the opti-mum pH at 6.0. The dechaining activity wascompletely destroyed by boiling the cell wallextracts for 30 min, suggesting typical enzymaticcharacter. It is of interest to note that theseextracts still showed considerable activity atpH's above 8.0 (Fig. 3), whereas only long-chained cell clumps were formed when the cellswere grown under such pH environments (Table1). This fact suggests that the formation of longchains of cells under alkaline conditions ismainly due to the difficulty in the synthesis ofthe enzyme rather than to its function. Thisassumption was supported by experimental re-sults shown in Table 2. The cell wall extractsfrom cells grown under various pH conditionsreacted under three different pH conditions. Ascan be seen, the extracts of cells grown at pH'sabove 8.0 did not show any dechaining activityeven at the optimum reaction pH of 6.0, sug-gesting that L. bulgaricus NLS-4 cells do not

J. BACTERIOL.

pH AND L. BULGARICUS CHAIN LENGTH

FIG. 2. Morphological differences of L. bulgaricus NLS-4 caused by the change of environmental pH. Allcells were observed microscopically when the continuous culture reached steady state after the shifts ofpHlevels. pH's were (A) 8.5; (B) 8.0; (C) 7.5; (D) 6.5; (E) 5.5; and (F) 4.5. Bar, 10 ,um.

I oo

501

v3 4 5 6

pH

TABLE 2. Effect ofgrowth reaction pH's ondechaining activity of cell wall extract

pH of Avg chain length (umJn)b after treatmentpHof ~~~~~atpH of-growth me-

dium' 4.0 6.0 8.0

4.5 100 40 706.0 90 10 607.5 100 60 808.0 100 100 1008.5 100 100 100

/ a Under these constant pH conditions, the cells werel grown continuously and then cell wall extracts were

prepared.b Cell chains with an average length of 100 lm,

obtained from the culture at pH 7.5, were treated withI I A the above cell wall extracts in the buffer as described

7 8 9 10 in Fig. 3.

FIG. 3. Heat treatment andpHofreaction solutioneffects on chain shortening activities of cell wallextracts. Cells having an average chain length of 100pum were treated with cell wall extracts for 2 h at37° C in reaction solutions having different pH's.Reaction solutions withpH's of3.0 to 5.0 are of 0.1 Mpotassium biphthalate-HCl buffer, pH's of 5.0 to 7.0are of 0.1 Mphosphate buffer, and pH's of 7.0 to 9.0are of 0.1 M Tris-HCl buffer. Symbols: *, fresh cellwall extracts from cells grown under pH 5.5, and A,the same extracts heated for 30 min at 100° C.

synthesize the dechaining enzyme under strongalkaline conditions.

DISCUSSIONWhen a continuous culture reaches steady

states, the metabolism of an organism in themedium can proceed under constant environ-mental factors independent of time. Continuousculture, therefore, is an ideal tool with which thetrue effect of environmental pH can be clarified

0xco2la

x

492

0

IL.0

867VOL. 144, 1980

868 RHEE AND PACK

because it eliminates effects of other variablefactors related to culture age usually accom-panying the conventional batch culture.The continuous culture experiments pre-

sented in this paper indicate close relationshipsbetween a medium pH and the chain length ofL. bulgaricus cells. Single or paired cells wereobserved only when the medium pH was as lowas 4.5, regardless of the growth rate. Dilutionrates, which are equivalent to specific growthrates at the steady state, did not affect the chainlength (data not shown). This finding contra-dicts the widely accepted belief that in thisspecies cells occur singly or in pairs when grownvigorously (11). The tendency of a filamentousgrowth under alkaline conditions was quite ap-parent. The reduction in chain length by me-chanical agitation, which has been reasonablyexpected (9), did not occur in this case. We triedincreased agitation rates without success (datanot shown); instead of chain-length reduction,the long threads of cells were curled, folded, androlled into tiny balls (Fig. 2A). Development ofvisible-sized balls, followed by washout, wereoccassionally observed under strong alkalineconditions.Involvement of autolytic enzymes in the sep-

aration of Pneumococcus (12) Streptococcus (3,8), Staphylococcus (1), and Bacillus (5, 6, 13,14) cells has been suggested. The present resultsindicate that autolytic activity is needed to de-chain L. bu4garicus cells and that environmentalpH determines the chain length of this organismby affecting the autolytic enzyme(s). The opti-mum pH for the dechaining activity of the en-zyme(s) was 6.0 (Fig. 3). However, when theorganism was grown under constant pH, theshortest chains appeared at a pH of -4.5 (Table1). Apparently, the mode of pH effect on thesynthesis and the activity of the enzyme(s) wasdifferent. This type of difference also has beenobserved in our laboratory with the lactate de-hydrogenase of the same organism; optimum pHfor activity was 6.5 and for synthesis was be-tween 5.0 and 6.5 (10). The synthesis of theautolytic enzyme(s) was completely suppressedat pH's above 8.0 (Table 2), whereas the alkaline

conditions did not inhibit the dechaining activitycompletely (Fig. 3). Therefore, it may be safelyconcluded that the filamentous growth of L.bulgaricus NLS-4 under alkaline conditions isdue to a failure in the synthesis of the enzyme(s)needed to dechain the cells.

LITERATU CITED1. Chatterjee, A. N., D. Mirelman, H. J.-Singer, and J.

T. Park. 1969. Properties of a novel pleiotropic bacte-riophage-resistant mutant of Staphylococcus aureus H.J. Bacteriol. 100:846-853.

2. de Vries, W., W. M. C. Kapteijn, E. G. van der Beek,and A. H. Stouthamer. 1970. Molar growth yields andfermentation balances of Lactobacillus casei L3 inbatch cultures and in continuous cultures. J. Gen. Mi-crobiol. 63:333-345.

3. Ekstedt, R. D., and G. IL Stollerman. 1960. Factorsaffecting the chain length of group A streptococci. I.Demonstration of a metabolically active chain-splittingsystem. J. Exp. Med. 112:671-686.

4. Elliker, P. R., A. W. Anderson, and G. Hannesson.1956. An agar culture medium for lactic acid strepto-cocci and lactobacilli. J. Dairy Sci. 39:1611-1612.

5. Fan, D. P. 1970. Cell wall binding properties of the Ba-ciUus subtilis autolysin(s). J. Bacteriol. 103:488493.

6. Fan, D. P. 1970. Autolysin(s) of Bacillus subtilis as de-chaining enzyme. J. Bacteriol. 103:494-499.

7. Koditechek, L K., D. Herdlin, and IL B. Woodruff.1949. Investigation on the nutrition of Lactobacilluslactis Doner. J. Biol. Chem. 179:1093-1102.

8. Lominski, I., J. Cameron, and G. Wyllie. 1958. Chain-ing and unchaining Streptococcus faecalis-a hypoth-esis of the mechanism of bacterial cell separation. Na-ture (London) 181:1477.

9. Martley, F. G. 1972. The effect of cell numbers in strep-tococcal chains on plate-counting. N. Z. J. Dairy Sci.Technol. 7:7-11.

10. Rhee, S. K., and M. Y. Pack. 1980. Effect of environ-mental pH on fermentation balance of Lactobacillusbulgaricus. J. Bacteriol. 144: 217-221.

11. Rogosa, M. 1974. Genus I. Lactobacillus Beijerinck 1901,212. Nom. cons. Opin. 38, Jud. Comm. 1971, p. 576-593.In E. Buchanan and N. E. Gibbons (ed.), Bergey'smanual of determinative bacteriology, 8th ed. The Wil-liams & Wilkins Co., Baltimore.

12. Tomasz, A. 1968. Biological consequences of the replace-ment of choline by ethanolamine in the cell wall ofpneumococcus; chain formation, loss of transformabil-ity, and loss of autolysis. Proc. Natl. Acad. Sci. U.S.A.59:86-93.

13. Young, F. E. 1966. Autolytic enzyme associated with cellwalls of Bacillus subtilis. J. Biol. Chem. 241:3462-3467.

14. Young, F. E., and J. Spizizen. 1963. Biochemical aspectsof competence in the Bacillus subtilis transformationsystem. II. Autolytic enzyme activity of cell walls. J.Biol. Chem. 238:3126-3130.

J. BACTERIOL.