Reprinted from the Proceedings of the NATIONAL ACADEMY OF SCIENCES,
Vol. 42, No. 9, pp. 574-577. September, 1956.

BACTERIAL PROTOPLASTS INDUCED BY PENICILLIN*

By Josaua LEDERBERG
DEPARTMENT OF GENETICS, UNIVERSITY OF WISCONSIN, MADISON, WISCONSIN

Communicated by Sewall Wright, May 28, 1956

Bacterial protoplasts are believed to be cellular units that have been deprived of
their rigid cell wall. Accordingly, they are distinguished by their spherical shape
(in bacilliform species) and their sensitivity to cytolysis in hypotonic media.
Several authors!~* have suggested that protoplasts might be useful material for the
study of biosynthesis of viruses and enzymes in a system simpler or more accessible
to external modification than the intact cell. So far, only gram-positive bacteria,
which are susceptible to lysozyme, have been used for such studies. A method has
now been found for the efficient production of protoplasts from enteric bacteria such
as Escherichia colt and Salmonella typhimurium, which are already familiar physio-
logical and genetic subjects. The technique consists essentially of the exposure of
growing cells to a medium containing penicillin, sucrose, and Mgt*.

Preparation of Protoplasts —Escherichia coli strain K-12 and a variety of its
mutant substrains are used in most of the experiments. The use of penicillin and
sucrose was suggested by the possible analogy between penicillin and lysozyme as
lytic agents and by the finding that hypertonic sucrose would interrupt bacterioly-
sis of Bacillus megatertum by lysozyme.' In addition, spherical bodies had been
casually noted in other applications of penicillin,‘ and many authors have em-
phasized its use in the production of L-forms.s The following procedure was
adopted after empirical trials and can doubtless be further improved.

The bacteria were grown overnight in tubes with 10 ml. of broth (Difco penassay
medium) at 37° C., ona rotator. Samples of 3 ml. of the grown culture (about 2
109 cells/ml) were added directly to 10 ml. of broth supplemented with penicillin,
1,000 u/mi, sucrose 20 per cent, and magnesium sulfate 0.2 per cent. In 2-3
hours the cells were quantitatively converted into spheres. During this interval
the optical density (measured at 650 my in a Coleman 14 spectrophotometer) in-
creased about 50 per cent, but the total count (spheres or rods, estimated in a
Petroff-Hausser chamber) remained constant. The spheres promptly lysed when
the suspension was diluted in distilled water, and they are therefore regarded as
“protoplasts.”

The indicated supplements are in substantial excess, and nearly optimal yields
of protoplasts can be obtained with 5 per cent sucrose, 100 u. penicillin, and 0.1
per cent MgSO,. The high magnesium requirement may depend partly on binding
with the sodium citrate used in the compounding of the penicillin preparation
Vou. 42, 1956 BIOCHEMISTRY: J. LEDERBERG 575

(Squibb penicillin G, potassium, buffered). The Mg++ can be replaced by Catt
but was preferred, to minimize precipitation of phosphate. In the absence of
Mg*+, protoplasts are formed but are less well preserved, and a third or less of the
initial rods are recovered. In the absence of sucrose or comparable stabilizer, the
cells are almost totally lysed, leaving only debris. In the absence of penicillin, the
rods grow at nearly the normal rate, i.e., are not appreciably inhibited by 20 per
cent sucrose.

Protoplasts were also formed in minimal medium, but the transformation has
been irregular and incomplete in the trials to date. Cells exposed to penicillin
under conditions not supporting growth, e.g., in sucrose buffer, were not trans-
formed into spheres.

Microscopic Observations and Viability —The transformation of rods into spheres
was observed in small droplets each containing ten to twenty cells, immersed in an
oil chamber.’ Each rod gave rise to a single sphere; even cells about to divide
were inhibited in further growth and division. The progression of stages as ob-
served with dark phase contrast at 645% is illustrated in Figure 1. A swelling
first appears, either subterminally or (in incipient division stages) centrally, and
progressively enlarges. The remainder of the rod then withers away and disap-
pears,

or aKkn

Fig. 1.—Above: Transformation of bacterial rod to spherical protoplast in the presence of peni-
eillin. Below: reversion of protoplast to rod in absence of penicillin, Observed under dark phase
contrast.

The reversion of spherical protoplasts to rods, in the absence of penicillin, has
been followed by a similar technique. The spheres first enlarge, then develop
amoeboid, finally rhizoid, outgrowths which segment to give typical rods. Spheres
incubated in the penicillin medium continue to enlarge but vacuolate and eventually
lyse over an interval of 24 hours. The optical density of the suspensions shows a
parallel cycle. The enlarging spheres are highly reminiscent of published figures of
L-forms,® but no evidence for multiplication of spherical elements has been seen so
far in these experiments or in trials in which 10 per cent bovine or equine serum was
furnished. The protoplast suspensions remain intact at least for several days at
4° C, After spontaneous lysis, or upon cytolysis in water, a residual ghost is seen
576 BIOCHEMISTRY: J. LEDERBERG Proc. N. A.S.

which retains its spherical form, although deprived of phase-dense material, The
lysates are quite viscous, presumably from the release of desoxyribonucleic acid.

About half the protoplasts are capable of reverting to rods and engendering
typical bacillary colonies in the absence of penicillin. This was verified by com-
parisons of total with viable counts on EMB lactose agar plus 20 per cent sucrose.
The viability of the protoplasts is completely lost by cytolysis in water. On the
other hand, normal rods are unimpaired when suspensions in sucrose solution are
similarly diluted.

Enzyme Synthesis and Other Observations——Protoplasts were tested for their
ability to synthesize 6-p-galactosidase: suspensions were sedimented, washed
with 20 per cent sucrose, and resuspended in lactose medium (casein digest. broth
plus sucrose plus MgSO, plus penicillin plus 1 per cent lactose). Over an interval
of several hours, samples were assayed with o-nitrophenyl galactoside.? In 210
minutes the enzyme activity increased from an initial value of 0.38 to 73 units per
10° spheres, a two-hundred-fold increase, while the optical density increased about
twofold. (One unit corresponds to the splitting of 10-8 moles of substrate per
minute. The peak activity found in spheres from lactose-grown cells is about 150
units. According to Rotman and Spiegelman,’ each of these units corresponds to
5 X 107° gm. nitrogen, or ~2 X 10” molecules of purified enzyme protein.)
Most of the induced enzyme activity sedimented with the protoplasts but was re-
leased to the supernate when the protoplasts were cytolysed in water.

Protoplasts were produced in a similar fashion in E. coli strain B and in S.
typhimurium strain TM2. The cells of the latter swam very actively, but the
protoplasts were immotile, except for an occasional sphere on which a bacillary
protuberance persisted (or had regenerated). Similarly, the protoplasts failed to
agglutinate in homologous antiflagellar serum. However, electron-microseopic
studies will be needed to establish the relationship of flagella to cell walls in this
species. Protoplast suspensions gave a faint agglutination in homologous somatic
antiserum.

The protoplasts of H. coli K-12 at first appeared to be readily stainable with
methylene blue but were later found to be very sensitive to photodynamic effects
which result either in ysis or stainability. Boiled protoplasts retain their form
and are evidently fixed as they become much more opaque, stain readily, and do not
cytolyse in water. Stained protoplasts often show a denser inclusion about one-
half their diameter, a poorly staining cortex, and a deeper-staining periphery. The
protoplasts reduce triphenyl-tetrazolium chloride, as well as methylene blue, very
readily. With the former, a typical formazan granule is produced, usually single
and near the periphery of the protoplast, much as in the rods. No corresponding
structure was seen in the untreated spheres.

In view of reports on the fusion of protoplasts,” and the possibility of DNA-
mediated transduction, preparations have been made with a variety of appropriate
genotypic mixtures of #. coli K-12. No evidence of hitherto unrecognized modes
of genetic recombination has been uncovered in preliminary trials. Protoplast
suspensions of appropriate compatibility types did, however, retain their ability to
mate.

Discussion.—Most of the findings in this preliminary survey warrant more in-
tensive and exact quantitative study. The areas of the interest already expressed
Vow. 42, 1956 BIOCHEMISTRY: J. LEDERBERG 577

in bacterial protoplasts! are already beyond the scope of any one laboratory, and a
ready method for their production in enteric bacteria may encourage further atten-
tion. In addition to their use in the extraction of intracellular constituents by
cytolysis and their availability for biosynthetic studies (enzymes, viruses, cell
walls), they may also be useful in cytological, antigenic, and physiological analysis.
Their possible genetic applications are currently being studied in this laboratory.

The preservation of viable protoplasts in sucrose solution supports the hypothe-
sis that the primary action of penicillin is on the synthesis or maintenance of a
component of the cell wall,!! while the activity and synthesis of other enzymes and
over-all protoplasmic mass are unaffected.

The protoplasts of #. coli show some instructive differences from those of B.
megaterium. The latter are rapidly produced by the enzymatic dissolution of the
existing wall, and each rod produces not one but several spheres. The latter dif-
ference presumably corresponds to the septate structure of gram-positive rods,
each compartment containing a single protoplast.’*7 Gram-negative bacteria, on
the other hand, are believed to be true multinucleate coenocytes. The most striking
difference may vanish with technical improvements—namely, that the protoplasts
of B. megaterium are reportedly inviable, while those of EZ. colt revert to normal rods,
an obvious advantage for genetic research.

Summary.—Growing cells of #. colt or S. typhimurium are quantitatively trans-
formed into spherical “protoplasts” in the presence of penicillin, magnesium, and
high concentrations of sucrose. The protoplasts are quickly cytolysed in water
but revert to viable rods in the absence of penicillin. The protoplasts have not
been observed to divide as such but have been shown to enlarge and to synthesize
an inducible enzyme (@-p-galactosidase) in appropriate media. The possible ap-
plications of these protoplasts in physiological and genetic research are cited.

* Paper No. 625 of the Department of Geneties. This work has been supported by grants
from the National Science Foundation and the National Cancer Institute, Public Health Service
(C-2157).

1k. T. C. Spooner and B. A. D. Stocker (eds.), Bacterial Anatomy: Symposium of the Society
for General Microbiology (Cambridge: At the University Press, 1956).

2, Landman and S. Spiegelman, these ProcrEeptines, 41, 698, 1955.

3§. Brenner and G. 8. Stent, Biochim. et biophys. acta, 17, 465, 1955.

4 J. Lederberg and N. D. Zinder, J. Am. Chem. Soc., 70, 4267, 1948; J. Lederberg, E. M. Leder-
berg, N. D. Zinder, and E. R. Lively, Cold Spring Harbor Symposia Quant. Biol., 16, 413, 1953.

5 L. Dienes and H. I. Weinberger, Bacteriol. Revs., 15, 245, 1951.

® J. Lederberg, J. Bactertol., 68, 258, 1954.

7 J. Lederberg, zbid., 60, 381, 1950.

8 B. Rotman and 8. Spiegelman, J. Bacteriol., 68, 419, 1954, and personal communication.

9 J. Lederberg, J. Bacteriol., 56, 695, 1948.

10 H. Stahelin, Schweiz. Z. allgem. Pathol. u. Bakteriol., 17, 296, 1954.

4 P. D. Cooper, Bacteriol. Revs., 20, 28, 1956.

12C, F. Robinow, in R. J. Dubos (ed.), The Bacterial Cell (Cambridge, Mass.: Harvard Uni-
versity Press, 1945).