Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

A CELL-MEMBRANE BOUND FRACTION OF BACTERIAL DNA
A. T. Ganesan and Joshua Lederberg
Department of Genetics
Stanford University School of Medicine
Palo Alto, California
Received January 25, 1965

In contrast to the abundance of information on the enzymatic
replication of DNA in vitro (Kornberg, 1962), the organization and
replication of the bacterial chromosome is not well understood. Genetic
and autoradiographic studies suggest a model of the bacterial genome as
a closed ring, only a small critical region of which is being replicated
at any given time (Sueoka, 1963; Cairns, 1963; Nagata, 1963; Lark and
Lark, 1964; Hanawalt and Ray, 1964). Jacob, et. al. (1963), proposed
a model in which replication might occur in direct association with
the cell membrane. There is no direct evidence for the site of DNA
replication in the cells, although Lark and Lark (1964) suggested the
possible involvement of some structural protein cell wall material in
the regulation of chromosome replication. However, many relevant facts
have been reported, mostly concerning the localization of nascent DNA,
i.e., DNA which is enriched in counts from a recent pulse of labeled
precussors,

Goldstein and Brown (1961) using E. coli protoplasts, showed the
occurrence of DNA synthesis in a sedimentable heavy particle fraction
which was resistant to sonic oscillation and rich in nascent DNA. Ben-
Porat, et. al. (1962), found that nascent DNA from mammalian cells was
preferentially agglomerated at an interface during deproteinization.
DNA-polymerase complexes in E. coli lysates were reported by Billen

(1962) and Kadoya, et. al, (1964). In B. subtilis, Ganesan (1963)

found the total DNA of cells could be pelleted after controlled lysis;

the complex had a molecular weight in the range of 50-70 million.

824
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Smith (1964), found that in E. coli lysed by detergent, 50% of nascent

DNA sedimented at low centrifugal speeds (2000 rpm) and was bound to

the particulate fraction, Electron micrographs of B. subtilis show
cellular DNA associated with cell membrane (Ryter and Jacob, 1963;

Ryter and Landman, 1964). Hanawalt and Ray (1964), found that nascent

DNA in E. coli was difficult to extract without treatment with proteolytic
enzymes. Detergent and phenol treatments discriminated against newly
replicated DNA. The nascent DNA containing fractions were pycnogra-~
phically less dense, suggesting a protein complex. A preliminary

account of the following results have been reported earlier (Ganesan

and Lederberg, 1965).

Materials and Methods
Stocks of labeled Precursors as purchased were:
Deoxythymidine-2-c!4, New England Nuclear,
30 mc/mM (0.807 mg/ml in H,0).
Deoxythymidine methy1-H°, New England Nuclear,
6.7 C/mM (0.18 mg/Sm1 in 4,0).
Deoxyadenosine triphosphate, we (NH,),) Schwartz BioResearch
1.2 C/mM,
A thymine requiring strain of B. subtilis (Thy Try ) (kindly given by
Dr. F. Rothman), which is kept in our stocks as SB 566, was used in
these experiments. A culture was grown for 3 generations (150 minutes

doubling time) in minimal medium! containing 10 ug/ml of DL~tryptophan

and 21 ug/ml of c!4_deoxythymidine having a specific activity of 8 x 107
CPM/ug in our counter. The DNA from such a culture at the end of 3 gener-
ations had a specific activity of 600 CPM/muM. The labeled culture was
washed two times and resuspended in thymine-free minimal medium at a

concentration of 4 x 10° cells/ml and incubated at 37°C. for 10 minutes
ee

1 Spizizen's minimal medium with 0.02% caesin hydrolysate (ref. 6).

825
’

if
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
¢

t

to deplete the internal thymine pool (no inviability was noted before
15 minutes). Deoxythymidine methyl-H> (.01 mc) was added to 2 x 10°
cells in 5 ml for different lengths of time. The growth was stopped
by addition of azide buffer!, 3x 10° cells in 30 A, having 2.14 mM
of DNA, were lysed with 50 ug of lysozyme followed by addition of 0.27
mi of the same buffer without EDTA. At this point, sodium dodecyl
sulfate was added to 0.1% concentration. The cleared lysate was
layered on a sucrose gradient (Burgi and Hershey, 1963), and spun at
36,000 rpm for 3 hours at 10°C. At the end of this treatment less
than 10> of the cells were viable. For polymerase assay, the same
procedure was used without detergent but with control of lysis under
the microscope. The lysis without detergent takes longer at 37°C. with
the same residual viability. Me’’ (1077 M) was essential for the
stability of sedimentable polymerase membrane complex, as well as for
its activity in the assay (Fig. 4). DNA polymerase was assayed
(Richardson, et. al., 1964), using dAT™ polymer as a primer with two
added triphosphates of deoxyadenosine and deoxythymidine. The dATP
H was diluted to give 1.5 x 10° epm/10 muM of dATP.

The backgrounds for the gradients were determined by running an

acid soluble supernatant in a sucrose gradient, followed by acid pre-

cipitation with added carrier. In general, a small proportion of acid
precipitable cl4 and we counts were found on the top of the gradient.
They were biologically inactive, presumably oligonucleotides. The
nascent DNA of cultures approaching the stationary phase is unstable
in lysates giving rise to acid soluble counts on standing. In the
reported pulse experiments, we have lost from 8 to 20% of the nascent

DNA, while the non-replicating DNA was completely recovered. (These

a

2

- - : -2
1 1072 M sodium azide, 107° M Tris-HCl, pH 7.8, 10°” M MgCl, 10° M

EDTA and 10-1 M NaCl. (0°C)

* copolymer of deoxyadenylate and deoxythymidylate.

826
Vol. 18, No. 5-6, 1965

AzmMrowMm VM

aAzmMoOwmMyD

aAzmMmoapame

 

   
 

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

 
    

 

 
    
 

 

 

 

Hg
20
Fig-4
15
TWO SECONDS

10

5

~PELLET- 3 10 15 20 25
FRACTIONS
20
H3
15 Fio-l
TEN SECONDS

10

5

-PELLET- 3 10 15 20 25
FARCTIONS
20
15. Fig.2
THIRTY SECQNDS

10 4 as]

 

C14 _/
J.

~PELLET- 5 10 15 20 25
FRACTIONS

827
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Figure lL.

Sucrose gradient sedimentations of lysates of pulse-labeled cells;
thymidine-H? added for 2, 10 and 30 seconds, respectively. The
pellet is represented as if it were evenly distributed among 4
fractions, Each fraction consists of 3 drops (.25 ml). The
distributions are normalized as percentages of the total counts,
as given below.

 

 

 

 

Time of Total Total muM muM of
Pulse H3CPM cl4cpm _of DNA New DNA
2" 223 1.2x10° 2.14 1.3x107"
10" 426 1.2«10° 2.14 2.6x107"
30" 876 1.2x10° 2.14 5.1x1074
Time of nox in co} % in w%in oly in
Pulse Pellet Pellet the Sup. the Sup.
2" 83.68 25.38 16.32 74.62
10" 65.30 23.07 34.7 76.93
30" 36.37 21.06 63.63 78.94

The same results were obtained using up to 5x the amount of DNA
and H3 pulse counts. In 8 experiments, the shortest pulse
always yielded from 68-96% of the nascent DNA in the pellet.

The above set is given because it was done at the same time with
the same cells and in the same gradient run, so all the condi-
tions are comparable with little variation. In these 8 experi-
ments, the bulk cl4 labeled non-replicating DNA varied from 20-
60% in the pellet. It is difficult to predict the amount of
non-replicating DNA in the pellet while the behavior of nascent
DNA is easily predicted. (Sup. refers to supernatant.)

The plot is computer-generated (on a Calcomp Digital Plotter
reading IBM 7090 output).

 

observations suggest either that the nascent DNA might be more suscep-
tible to nucleases or perhaps more likely, that the nucleases are also
localized in the sedimentable fractions.)

Differential counting was done using a Packard TriCarb Liquid
Scintillation Counter. In general, low amounts of lysate gave very
reproducible results with no aggregation effects. Transformation

assays were performed as described by Ganesan and Lederberg (1964).

828
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Results
Figure 1 shows that the nascent DNA (He labeled), is predominantly
associated with the pellet fraction. Electron microscope studies
indicated predominantly membranes and broken cell wall structures in
the pellet. Figure 2 shows the uw counts incorporated during different

lengths of "pulses", and suggests a constant rate of DNA synthesis
after the first datum (2 seconds). The initial rate may show an error
in the time base of 9 seconds, perhaps due to delay in stopping the
growth. The absolute amount of non-nascent DNA found in the pellet
varied considerably from experiment to experiment. If replication is
localized, then the separation of nascent DNA from the chromosome

depends on the distribution of breaks which depends on capricious

fluid shear. Very careful lysis with no agitation and no pipetting

gives lysates from which all the DNA sediments out as pellet.

Figure 2. Incorporation of we thymidine by 3 x 10° cells.

cpm

Boo

The dotted line represents
the counts of "pulsed" DNA
that appeared in the super-
s natant of the gradients

“7 (sup.), while the solid
line represents the total
counts in the DNA (2.14
muM).

 

 

) 10 20 50
TIME IN SECONDS

 

Microscopic examination during the process of lysis reveals that
lysozyme attacks the cell wall at irregular intervals. The cells
become fragile and in the absence of stabilizing agents the bulk of

the cytoplasm and some RNA leaches out of the ruptured membrane. We

829
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

interpret this to mean that the DNA still attached to the membrane is
protected by it and sediments easily. This might be the mechanism of
"nuclear body" formation which was observed by Spiegelman, et. al.

(1958). The chemical composition of this pellet (Ganesan, 1963), is
similar to that found in B. megatherium by Spiegelman, et. al. (1958)

and Butler and Godson (1963).

Chase Experiments

If the replicating point(s) were attached to a specific site on

the membrane cell wall, it should be possible to chase the “pulse” label

 

 

 

 

20
e CHASED IN THYMIDINE MEDIUM
E Fig.3
A
c 15 TEN SECONDS
E
N Hg
T 104
ci4
5
-PELLET- 5 10 15 20 25
FRACTIONS
20
CHASED IN Ci4 THYMIDINE MEDIUM
P Fie.3
E
Ais TEN SECONDS
c
E
N
T 10
Ci¥
5 |
H3
-PELLET- 5 10 15 20 25

FAACTIONS

830
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Figure 3.

uw pulse chased in ci4 thymidine medium.

Pellet Sup.
cl4 24% 76%
H? Pulse 9.06% 90.94%
(x? 100% = 4.6 x 10° cpm while ct“ = 1 x 104 cpm.)

H pulse chased out in cold thymidine medium.

 

ci4 43% 57%
Ww 42.4% 57.6%
ca? 100% = 8.7 x 10° cpm while cl4 = 1.7 x 102 cpM.)

These experiments were also tried with different amounts of DNA
per gradient. More than 10 myM of DNA in a 5 ml gradient (approx.
2 x 109 cell equivalents) leads to aggregation so that bulk of
DNA sediments out as a gel in the pellet.

 

from the pellet by subsequent growth in cold thymidine, as indicated
by a uniform ratio of wycl4 constant throughout the gradient. If the
uw pulse were chased by subsequent growth in cl4 thymidine medium, the
ratio of wycl4 should change drastically, especially in the pellet.

These results are shown in Figure 3 for 10" pulse.

DNA Polymerase Activity

Cells labeled uniformly with ci4 thymidine were used in this experi-

ment (Figure 4). The DNA polymerase complex is Mg’? dependent. The
pellet contains 1/3 of the DNA. The fractions most active in polymerase
are in the pellet and top of the gradient. Treatment of the pellet with
pronase or omitting one of triphosphates or ugtt resulted in loss of
activity, Under these conditions the cellular DNA did not have any

inhibitory effect on the synthetic ability with dAT ‘as primer. The

831
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

      

3 Vy
cpm. H “Mg 1h cpm.H hk cpm. C
cpm. C 10M. Mg
POLYMERAS e 600 6 600
00 .
5000 POLYHERMSE

  

Fig. 4
7 3) :
| 20% 14) POLYMERASE ACTIVITY (17? oo 4000
(c'*? oNA o----8 ° a
fe cpm oem. Mg (cpm. H7)
{or 30h
oN
y 3000 7
i
\
PELLET A

 

 

 

 

5 15 15 2 50 FRACTIONS
Figure 4.

DNA polymerase complexes. The gradients were run with 10+

M $-mercaptoethanol. Mgt+ was added as indicated. The inserts
show the effect of Mgtt at O and 10-4 M while the main figure
shows the effect at 10-2 M concentrations. Note the pellet has
no enzymatic activity or appreciable amount of DNA at 0 and 10-4
M Mg ++ concentrations. In all these cases residual unlysed
cells were removed by centrifugation prior to zone sedimenta-
tion. 100% corresponds to 3 x 104 CPM per 5 mumoles of DNA and
12 ug of protein was added per gradient. In the insert figures,
the specific activity of DNA was less while the concentrations

are approximately the same. 50 A of each fraction was assayed
for polymerase.

 

results of detailed kinetic experiments are under publication (Ganesan,
et. al., 1965). From these experiments we conclude that the membrane

fraction possesses up to 25% of the total assayable DNA polymerase

activity.

832
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Discussion
B. subtilis has a genome of 1.8 x 10° nucleotide pairs. Under
these conditions with a generation time of 150 minutes, 1.2 x 104

nucleotide pairs replicates every minute. When the cells are washed

and starved as in these "pulse" experiments, the replication is very
slow. The slope of figure 2 indicates a rate of the order of 5.7 times
slower than the normal replication rate. This may be due to the washing
procedures introduced before exposure to we thymidine. The nascent DNA,
after 2 seconds exposure to Hw thymidine, bands in CsCl gradient at a
stratum of 1.703 gms/ec.

As far as can be judged from the limited data of figure 2, i.e.,
the displacement of the 2 curves, the transit time of nascent DNA ts
about 10.5 seconds, corresponding to 168-200 nucleotides. Here we
assume all the nascent DNA is formed from the tracer precursors
supplied. Nascency is defined operationally by a fragmentation process;
hence this estimate is, if anything, only the upper limit of any real
unit in the cell.

The site of attachment of the "replicator" seems to be of protein
nature. Treatment of the lysate before detergent treatment with pronase
releases the nascent DNA which bands at a higher stratum in a sucrose
gradient compared to the non-replicating DNA. The patterns are consis-—
tent with the length of exposure to we thymidine. The shift in sedimen-
tation is assumed to be due to the dissociation of the replicating
point from the membrane so that the nascent DNA is of smaller molecular
size compared to the old DNA. This is based on the assumption that the
results obtained are not due to any configurational differences between
nascent and bulk DNA to impose the observed differential mobility in a
sucrose gradient. The molecular weights were calculated using TS DNA
as standard in sucrose zone sedimentation. In E. coli the nascent DNA,
enriched in the pellet, dissociates on subsequent treatment with sodium

deoxycholate (Smith and Hanawalt, 1965).

833
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

In experiments of pulse labeling with uw uridine for 30 seconds,
25-30% of the "messenger RNA" of the cells was found in the pellet.
Several authors have shown that protein synthesis occurs on the mem-
brane of bacterial cell (see Schlessinger, 1963). In our experiments
the pellet fraction was found to incorporate all 4 deoxynucleoside
triphosphates using the primer already present. This is Met? dependent
and requires all 4 triphosphates. The details of all these experiments
mentioned above will be published elsewhere,

From the above observations, the most plausible explanation is
that membrane cell wall is the site of DNA synthesis in bacteria,

The discouraging results of efforts (Richardson, et. al., 1963,
etc.) to demonstrate substantial net replication of DNA with intact
genetic activity through the in vitro system can probably be explained
by the interruption of coherent linear replication. This results in
the formation of DNA whose thermal denaturation is spontaneously rever-
sible and which appears to have a highly involuted structure (Schildkraut,
et. al., 1964). This has raised some question whether the DNA poly-
merase now studied in vitro is the crucial enzyme of genetic replication;
however, its association with organized structures may provide the
essential basis of continued orderly replication.

We are delighted to acknowledge innumerable points of guidance
and support from Professor Arthur Kornberg and his colleagues of the
Department of Biochemistry, and the skillful technical assistance of
Mrs. Susan Grether.

These studies have been supported in part by National Institutes
of Health Training Grant 2TI-GM-295 and Research Grant AI-5160, and by

the National Science Foundation Research Grant G-6411.

REFERENCES
Ben-Porat, T., A. Steere, and A. S. Kaplan, Biochim. Biophys. Acta,

61, 150 (1962).
Billen, D. Biochem. Biophy. Res. Comm. 3, 179 (1962).

834
Vol. 18, No. 5-6, 1965 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Burgi, E. and A. D. Hershey, Biophys, J. 3, 309 (1963).

Butler, J. A. V. and G. N. Godson, Biochim. J. 88, 176 (1963).

Cairns, J. Cold Sp. Harb. Sym. Quant. Biol.: XXVIII 329 (1963).

Ganesan, A. T., Ph.D. Thesis, Stanford University (1963).

Ganesan, A. T., S. Grether and J, Lederberg (under prep.).

Ganesan, A. T. and J. Lederberg, J. Mol. Biol. 9, 683 (1964).

Ganesan, A. T. and J. Lederberg, Biophysical Society-Abstracts (1965).

Goldstein, A. and B. J. Brown, Biochim. Biophys. Acta, 53, 19 (1961).

Hanawalt, P. L. and D. S. Ray, Proc. Nat'l. Acad. Sci., U. S., 52,
125 (1964).

Jacob, F., S. Brenner and F. Cuzin, Cold. Sp. Harb. Sym. for Quant.
Biol. XXVIII, 329 (1963).

Kadoya, M., H. Mitsui, Y. Takagi, FE, Otaka, H. Suzuki, and S. Osawa,
Biochim. Biophys. Acta 91, 36 (1964).

Kornberg, A. Enzymatic Synthesis of DNA. Wiley, N. Y. (1962).

Lark, C. and K. G. Lark, J. Mol. Biol. 10, 120 (1964).

Nagata, T., Cold Sp. Harb. Sym. for Quant. Biol. XXVIII, 55 (1963).

Richardson, C. C., C. L. Schildkraut, H. V. Aposhian, A. Kornberg,
W. Bodmer and J. Lederberg. Studies on the Replication of DNA
by Escherichi coli Polymerase. Symp. on Informational Macro-
molecules, Academic Press, N. Y. (1963).

Richardson, C. C., C. L. Schildkraut, H. V. Aposhian and A. Kornberg,
J. Biol. Chem., 239, 222 (1964).

Ryter, A. and F. Jacob, Gompt. Rend. 257, 3060 (1963).

Ryter, A. and 0. E. Landman, J. Bact. 88, 457 (1964).

Schildkraut, C. L., C. C. Richardson and A. Kornberg, J. Mol. Biol.
9, 24 (1964).

Schlessinger, D., J. Mol. Biol. 7, 569 (1963).

Smith, K. C., Personal communication.

Smith, D. W. and P. C. Hanawalt, Biophysical Society-Abstracts (1965).

Spiegelman, S., A. I. Aronson, and P. C. Fitz~James, J. Bact. 75,
102 (1958).

Sueoka, N., and H. Yoshikawa, Cold Sp. Harb. Sym. for Quant. Biol.
XXVIII, 47 (1963).

835