bacteriovorus 109J - Journal of Bacteriology

JOURNAL OF BACTERIOLOGY, Nov. 1979, p. 620-633
0021-9193/79/11-0620/14$02.00/0
Vol. 140, No. 5
Regulated Breakdown of Escherichia coli Deoxyribonucleic
Acid During Intraperiplasmic Growth of Bdellovibrio
bacteriovorus 109J
REINHARDT A. ROSSONt AND SYDNEY C. RITTENBERG*
Department of Microbiology, University of California, Los Angeles, California 90024
Received for publication 3 July 1979
The nucleic acid metabolism of Bdellovibrio
bacteriovorus during intraperiplasmic growth
includes the degradation of the substrate organism's nucleic acids and the synthesis of its own
nucleic acids (15, 16, 25). These processes appear
to be highly regulated. The degradation of the
DNA and RNA of the substrate organism into
large fragments is apparently completed prior to
the initiation of bulk synthesis of bdellovibrio
nucleic acids. The bdellovibrio preferentially utilizes the degraded nucleic acid fragments as precursors for its own DNA and RNA synthesis (15,
16, 25). We have previously shown (25) that in
the first 45 to 60 min of a 3- to 4-h growth cycle,
the DNA of a substrate organism is rapidly and
completely degraded to pieces that are largely
cold-acid-insoluble and retained within the bdelloplast. The bdellovibrio then initiates its own
DNA synthesis. At no time during its growth is
there an appreciable rate of release of acid-soluble DNA products from the bdelloplast. This
suggests that the rates of breakdown of the DNA
fragments to soluble pieces and the rates of
bdellovibrio uptake and polymerization of these
DNA products are similar. At completion of
t Present address: Marine Biology Research Division,
Scripps Institution of Oceanography, University of California,
La Jolla, CA 92093.
620
bdellovibrio development on a cell of normal
composition, about 30% of the initial DNA of
the substrate organism is soluble and the remaining 70% is incorporated into bdellovibrio
DNA (25).
The highly regulated nature of DNA degradation suggests that bdellovibrio enzymes are
responsible for the process. However, the possibility that DNases of the substrate cell function
exclusively or in concert with bdellovibrio enzymes for DNA breakdown must also be considered. In this paper we present data on DNase
activity and the kinetics of substrate cell DNA
breakdown during intraperiplasmic growth of B.
bacteriovorus on normal and heat-treated Escherichia coli. In addition, the effects of protein
synthesis inhibitors on these processes were examined. The results of these investigations lead
to the conclusion that the regulated degradation
of the DNA of the substrate organism during
development of B. bacteriovorus is a consequence of sequential synthesis and activity of
DNases of the bdellovibrio.
MATERIALS AND METHODS
Growth and harvesting of cells. B. bacteriovorus 109J was the experimental organism, and normal
or heated [2-'4C]deoxythymidine-labeled E. coli ML35
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
During growth of Bdellovibrio bacteriovorus on [2-"C]deoxythymidine-labeled
Escherichia coli, approximately 30% of the radioactivity was released to the
culture fluid as nucleoside monophosphates and free bases; the remainder was
incorporated by the bdellovibrio. By 60 min after bdellovibrio attack, when only
10% of the E. coli deoxyribonucleic acid (DNA) had been solubilized, the substrate
cell DNA was degraded to 5 x 105-dalton fragments retained within the bdelloplast. Kinetic studies showed these fragments were formed as the result of
sequential accumulation of single- and then double-strand cuts. DNA fragments
between 2 x 103 and 5 x 105 daltons were never observed. Chloramphenicol,
added at various times after initiation of bdellovibrio intraperiplasmic growth on
normal or on heated E. coli, which have inactivated deoxyribonucleases, inhibited
further breakdown and solubilization of substrate cell DNA. Analysis of these
intraperiplasmic culture deoxyribonuclease activities showed that bdellovibrio
deoxyribonucleases are synthesized while E. coli nucleases are inactivated. It is
concluded that continuous and sequential synthesis of bdellovibrio deoxyribonucleases of apparently differing specificities is necessary for complete breakdown
and solubilization of substrate cell DNA, and that substrate cell deoxyribonucleases are not involved in any significant way in the degradation process.
VOL. 140, 1979
DNA BREAKDOWN DURING B. BACTERIOVORUS GROWTH
T4D and 429 phages labeled with [methyl-3H]deoxythymidine in their DNAs were prepared by modifications of standard phage procedures (1, 28) and used as
sedimentation markers. The phage suspensions were
stored over CHC13 at 40C.
Lysis. Frozen culture samples were thawed at room
temperature, and either labeled T4D (2,500 cpm) or
4029 (2,500 cpm) was added. Samples (0.1 to 0.2 ml) for
neutral gradients were lysed in a total volume of 0.6
ml containing (final concentration): 0.15 M NaCl; 0.1
M EDTA (pH 8.5); sodium dodecyl sulfate, 8.5 mg/ml;
and proteinase K (Beckman Instruments Inc., Palo
Alto, Calif.), 350 jug/ml. The mixtures were incubated
for 2.0 to 2.5 h at 45°C. Samples (0.1 to 0.2 ml) for
alkaline gradients were lysed in a total volume of 0.7
ml containing (final concentrations): 0.5 M NaCl; 0.1
M EDTA (pH 8.5); sodium dodecyl sulfate, 15 mg/ml;
and proteinase K, 300 ,g/ml. The mixtures were incubated for 2 h at 450C. Sodium hydroxide (4.5 N) was
added to give a final concentration of 0.1 N, and the
samples were incubated for an additional 30 min at
room temperature. The alkaline samples had a measured final pH of 12.4 (unadjusted for Na+ concentration). All samples were mixed after additions and at
periodic intervals during incubations by gentle rolling
to minimize shear forces.
Gradient preparation. Linear 5 to 20% (wt/wt)
sucrose gradients were prepared by a modification of
the method of Barth and Grinter as described by
Jacob and Hobbs (20). Buffers for neutral and alkaline
sucrose solutions contained 0.15 M NaCl and 0.1 M
EDTA (pH 8.5) or 0.5 M NaCl, 0.1 M EDTA, and 0.1
N NaOH (pH 12.4), respectively.
To prepare neutral gradients, 4.9 ml of a 12.5% (wt/
wt) sucrose solution was added to Beckman SW50.1
cellulose nitrate centrifuge tubes. A 60% (wt/wt) sucrose solution (0.2 ml) was injected by syringe below
the 12.5% sucrose column. These tubes were then
frozen and stored at -70°C. To form the gradient,
tubes were thawed for 2 h in iced water and then
equilibrated to room temperature over an additional
30 to 60 min. At this point, a linear 5 to 20% gradient
over a high-density sucrose shelf was established.
Analysis showed that the gradient was linear to within
0.9 ml of the tube bottom (approximately the bottom
10 fractions).
Alkaline gradients were prepared in Beckman
SW50.1 polyallomar centrifuge tubes filled with 4.8 ml
of 12.5% (wt/wt) sucrose and 0.1 ml of 60% (wt/wt)
sucrose in alkaline buffer as described for neutral
gradients.
Gradient sedimentation: collection and analysis of samples. Lysed cultures, 0.1 to 0.2 ml of
sample, were layered on a gradient. Neutral and alkaline gradients contained 5.5 to 11.0,ug and 1.8 to 3.7
,ug of DNA, respectively. Approximately 60% of the
DNA was labeled E. coli DNA; the remainder was
unlabeled B. bacteriovorus DNA. The gradients were
centrifuged in a Beckman SW50.1 rotor at centrifugal
forces and times appropriate to the anticipated size of
the degraded E. coli DNA in the sample. Specific
details are listed in the legends to the figures.
Gradients were collected from the bottom through
a needle used to puncture the centrifuge tube. A
gradient was fractionated (49 to 50 equal fractions)
onto Whatmann 3MM disks (2.3 cm) by pumping with
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
(lacI lacY) served as the substrate for its intraperiplasmic growth. To obtain inocula for single-cycle
growth experiments, bdellovibrios were cultured on E.
coli cells (38) suspended in dilute nutrient broth (34).
Unless indicated otherwise, E. coli was grown in nutrient broth at 300C as previously described (31). To
obtain heat-treated cells, overnight (16-h) E. coli cultures, about 5 x 10" cells per ml, were rapidly warmed
to and incubated at 540C for 20 min with periodic
shaking. The heated cultures were then immediately
cooled to 4°C and held at this temperature until
harvesting. E. coli celis labeled with [2-14C]deoxythymidine were prepared by two modifications of the
method of Boyce and Setlow (6). With either method,
greater than 95% of the radioactive label incorporated
was in the nucleic acid fraction of the cells.
Method A. A 16-h E. coli culture was diluted to 108
cells per ml with fresh medium and incubated for 45
to 60 min on a rotary shaker at 370C. Deoxyadenosine
(750 ug/ml) and [2-"4C]deoxythymidine (specific activity, 58.5 mCi/mmol; 0.2 ,tCi/ml) were then added.
Incubation was continued for an additional 90 mi,
which permitted approximately 3.5 cell divisions and
a yield of about 109 log-phase cells per ml.
Method B. Nutrient broth was supplemented (per
ml) with 1 mg of deoxyadenosine and 0.2 ,uCi of [2'4C]deoxythymidine (specific activity, 58.2 mCi/
mmol). The medium was inoculated with E. coli (10'
cells per ml) and incubated on a rotary shaker at 300C
for 16 h, to give 5 x 109 to 7 x 109 stationary-phase
cells per ml.
B. bacteriovorus and E. coli cells were harvested
and washed two times with 0.01 M N-2-hydroxyethylpiperazine-N'-2'-ethanesulfonic acid (HEPES) buffer
(pH 7.6) by centrifugation at 4°C for 10 min at 10,000
x g. Bdellovibrio cultures were centrifuged for 2.5 min
at 1,000 x g before harvesting, and the pellets were
discarded to remove unlysed E. coli cells and debris.
The washed cells were resuspended in 0.01 M HEPES
buffer (pH 7.6) containing 3 x 10-3 M CaCl2 and 3 x
l0-4 M MgCl2 (HM buffer). Cell numbers in suspensions were determined from turbidity measurements
by reference to standard curves based on plaque
counts or colony counts (29, 40).
Growth experiments. Suspensions of B. bacteriovorus and [2-'4C]deoxythymidine-labeled E. coli in
HM buffer were mixed (0 time) to initiate experiments.
Initial E. coli populations of 5 x 109 cells per ml and
input ratios (B. bacteriovorus to E. coli cells) of 1.5 to
2.0 were used in all experiments. This resulted in a
rapid attack on all substrate cells and synchronous
single-cycle growth of the bdellovibrios. The bdellovibrio cultures were incubated with shaking at 310C
in a water bath. In some experiments, portions of a
culture were removed at various times, treated with
chloramphenicol (200 ,g/ml), puromycin (500 ug/ml),
or rifampin (100 tig/ml), and shaken in parallel with
the untreated culture. The cultures were sampled at
appropriate times to follow kinetics of DNA degradation. Samples for gradients were immediately frozen
in a Dry Ice-ethanol bath and stored at -70°C until
analysis. All other samples were analyzed immediately.
Bacteriophage. The E. coli bacteriophage T4D
was supplied by F. A. Eiserling, and B. subtilis bacteriophage 429 was obtained from J. Spizizen. Purified
621
622
ROSSON AND RITTENBERG
added to all samples before treatment. Sonication was
at maximum output for 1.5 to 2.0 min (50 to 60% duty
cycle) with a Branson Sonifier, model W200P, fitted
with a micro-tip probe. Typically, two to three sonic
treatments of 2-min duration were required to obtain
99% breakage, as estimated by microscopic examination. Sample temperature was maintained below 40C
with a Dry Ice-ethanol bath. After sonication, the
glass beads and unbroken cells were removed by centrifugation (2.5 min, 1,000 x g) at 40C and discarded.
Extracts that were not immediately analyzed were
rapidly frozen in a Dry Ice-ethanol bath and stored at
-70°C until used.
Purified [3H]DNA. [Methyl-3H]deoxythymidinelabeled DNA, extracted from E. coli, was the substrate
for all enzyme assays. Labeled E. coli cells were obtained by procedure A (see above), substituting glucose-mineral salts medium supplemented with 0.1%
Casamino Acids (17) for nutrient broth. The DNA of
the labeled cells was extracted by the method of
Marnur (24). The isolated DNAs, which had optical
density at 260 nm/optical density at 280 nm ratios
between 1.91 and 1.93 and specific activities of 9.0 x
103 to 4.6 x 104 cpm/yg, were stored frozen at -70°C
until used.
Enzyme a88says. Assays for endonuclease I and
exonucleases I, II, and III were modifications of the
procedures of Shortman and Lehman (35). To determine endonuclease I, extracts were treated with
RNase (100,ug/ml) or tRNA (100 ug/ml), then warmed
to and incubated at 370C for 10 min. Treated extracts
were added to 0.3 ml of assay mixture (35) and incubated at 370C for 30 min. Ice-cold 1.5 N HC104 (0.3
ml) and carrier calf thymus DNA (1.5 mg/ml; 0.2 ml)
were added, the mixtures were held on ice for a minimum of 15 min, and then the cold-acid-insoluble material was removed by centrifugation (15 min, 15,000
x g) at 40C. Radioactivity of 0.5-ml samples of the
supernatant was determined. Exonuclease contribution (tRNA-treated sample activity) was subtracted
from the RNase-treated sample activity. To assay for
exonucleases, extracts were added to appropriate assay
mixtures (35) containing native or heat-denatured
[3H]DNA and incubated at 37°C, and samples were
taken, treated, and analyzed as for endonuclease I.
To compare nuclease activity of untreated and
heated E. coli, [2-14C]deoxythymidine-labeled cells
were used, and the rate of formation of cold-trichloroacetic acid-soluble radioactivity from the endogenous substrate was measured in crude extracts.
The total potential DNase activity of bdellovibrio
cultures was approximated by measuring the rate of
release of cold-trichloroacetic acid-soluble radioactivity from exogenous DNA by crude extracts (i.e., the
sonicated bdelloplasts). The extracts (0.75 ml) were
supplemented with purified [3H]DNA (0.25 ml) and
warmed to 31°C. The added DNA comprised about
15% of the total extract DNA present (purified DNA
added plus sonicated bdellovibrio culture DNA). Duplicate samples (0.15 ml) were removed at 0 and 30
min and mixed with an equal volume of ice-cold 10%
(wt/vol) trichloroacetic acid. The acidified samples
were held at 40C overnight. Cold-acid-insoluble material was removed by filtration through 0.22-jm mem-
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
an LKB peristaltic pump at a constant flow rate. The
disks, while still wet, were immersed in 150 ml of icecold 5% (wt/vol) trichloroacetic acid and held on ice
for 30 min. They were then washed (100 ml per wash)
sequentially with ice-cold 5% trichloroacetic acid (one
time), 95% ethanol (two times), and anhydrous diethyl
ether (one time) and air dried (4), and their radioactivity was counted.
With the exception of gradients made from the 30min samples, essentially all radioactivity added to the
gradients was recovered (see Table 3). For reasons not
understood, it was difficult to get complete lysis of the
30-min samples, and the material layered on the gradients was always faintly turbid. It is possible that the
low recoveries from the 30-min samples were due to
the sedimentation and adherence of the nonlysed material to the bottom of the centrifuge tube.
The average molecular weight ofbandable substrate
cell DNA was determined by the equation of Burgi
and Hershey (7): D1/D2 = (Ml/M2)a, where Di, Ml, D2,
and M2 are the relative sedimentation distances and
molecular weights of two DNAs, respectively. The
sedimentation distance of a DNA band was calculated
from its peak center relative to the gradient meniscus.
Values for a are those derived by Studier (37) for
neutral (0.345) and alkaline-denatured (0.400) DNA.
These values apply for DNA over a molecular weight
range from below 106 to above 108. Sedimentation
marker DNA molecular weights are 1.25 x 10' for
bacteriophage T4D (10) and 107 for bacteriophage $29
(3).
Cold trichloroacetic acid-soluble radioactivity: preparation, fractionation, and analysis.
Samples of bdellovibrio cultures (2 ml) were immediately mixed with 1.0 ml of ice-cold 15% (wt/vol) trichloroacetic acid and held on ice for 15 min. The
mixtures were centrifuged for 15 min at 15,000 x g at
40C. Portions of the supernatant fluid (0.1 to 0.5 ml)
were used for determination of radioactivity. The trichloroacetic acid in the remaining supernatant fluid
was removed from the aqueous phase by extraction
with diethyl ether (2 volumes, six consecutive extractions). The aqueous phase was then lyophilized to
dryness and stored at -70°C until analysis. For analysis, a lyophlized sample was dissolved in 0.5 ml of 50
mM ammonium acetate buffer (pH 6.8), and 0.1 ml of
this solution was used to measure radioactivity. Then
0.3 ml was layered on a Biogel P2 column (0.7 by 60
cm), and the sample was eluted with 50 mM ammonium acetate buffer. Fractions (0.5 ml) of the eluant
were collected. The elution order of the cold-trichloroacetic acid-soluble components, established with
standards, was: oligonucleotides (greater than 1,800
daltons) at the exclusion volume, then dTMP, deoxythymidine, and thymine. Radioactivity in a fraction
was determined by counting 50 pl of sample.
Extract preparation. Crude cell-free extracts
were made by sonicating 2.5 ml of cold (400) B.
bacteriovorus or E. coli cell suspensions (0.1 g [wet
weight] of cells per ml of HM buffer) or bdellovibrio
cultures (10'0 B. bacteriovorus and 5 x 109 E. coli cells
per ml) supplemented with phenylmethylsulfonylfluoride (50 ug/ml). Glass beads (5 plm, ca. 0.2 ml/ml;
Branson Ultrasonic Systems, Danbury, Conn.) were
J. BACTERIOL.
DNA BREAKDOWN DURING B. BACTERIOVORUS GROWTH
VOL. 140, 1979
RESULTS
E. coli DNA degradation products formed
during B. bacteriovorus growth. The amount
and nature of cold-trichloroacetic acid-soluble
products generated during intraperiplasmic
growth was determined by analysis of samples
taken at hourly intervals from synchronous single-cycle cultures of B. bacteriovorus growing
on [2-'4C]deoxythymidine-labeled E. coli. In a
typical experiment the total cold-trichloroacetic
acid-soluble radioactivity increased slowly during bdellovibrio development, amounting to less
than 10% at 60 min, and reached a maximum of
32% at lysis (Table 1). The remaining 68% was
incorporated into the progeny bdellovibrio cells.
If the samples were separated into bdelloplast
(sedimentable) and supernatant fractions by
centrifugation, essentially all radioactivity soluble in trichloroacetic acid was in the supematant
fluid and all trichloroacetic acid-insoluble radio-
activity was in the pellet; that is, the soluble
DNA products equilibrated with the suspending
fluid, while the large DNA fragments did not
escape from the confines of the bdelloplast structure. These results are in general agreement with
those previously reported (25).
Standard gel sieving techniques using Bio-Gel
P2 showed (Table 1) that the acid-soluble radioactive material was primarily monomeric and
was distributed between dTMP, deoxythymidine, and thymine. An oligonucleotide fraction
slowly and steadily increased its relative proportion to the total soluble material to a maximum
of less than 25%. This material eluted just after
the exclusion volume, which suggested an average molecular weight of a little less than 1,800.
Between 60 and 180 min there was no significant change in the relative proportions of the
monomers. Thymine and dTMP were present in
the highest percentage, whereas deoxythymidine comprised less than 5% of the cold-trichloroacetic acid-soluble material. All dTMP fractions were completely hydrolyzed to deoxythymidine and inorganic phosphate by 5'-nucleotidase, thus showing that dTMP was as deoxythymidine 5'-monophosphate (27). Between 180 and
240 min there was a dramatic decrease in both
the total and relative amounts of dTMP radioactivity and concomitant increases in both the
deoxythymidine and thymine fractions. Apparently a change in the nature of the pyrimidine
degradative enzymes occurred after 180 min, a
time corresponding to the completion of bdellovibrio development and the onset of lysis of the
bdelloplasts. The observed distribution of radioactivity in the cold-trichloroacetic acid-soluble
radioactive material at the completion of bdellovibrio growth in these experiments parallels that
found in culture supematants of bdellovibrios
grown on [2-14C]uracil-labeled E. coli (15).
The nature of the cold-trichloroacetic acid-
TABLE 1. Release and nature of cold-acid-soluble E. coli DNA products during intraperiplasmic growth of
B. bacteriovorus on [2-'4C]deoxythymidine-labeled E. colia
Sample
Total acid-soluble radioactiv-
time
ity
Radioactivity in:
Oligonucleo-
dTMP
TdR
T
% Recovery
%b
%c cpm/ml %C cpm/ml %C cpm/ml %C
cpm/ml
850
4.4 10,500 54.1
98
6,370 32.7
1,330 6.8
19,500 4.8
1.0 15,000 40.9
380
104
4,750 13.3 17,700 48.2
36,600 9.0
99
3,810 4.6 33,800 40.9
82,500 20.3 13,700 16.6 28,300 34.3
102
3,000 2.3 20,700 16.1 77,200 60.1
128,000 31.6 29,900 23.3
a
The culture initially contained 7.5 x 109 B. bacteriovorus and 5 x 109 E. coli cells per ml (406,600 cpm).
Samples were taken at indicated times after mixing.
b As percentage of total radioactivity added initially.
As percentage of total radioactivity in cold-trichloroacetic acid-soluble fraction (column 2). TdR, Deoxythy-
cpm/rnl
60
120
180
240
c
midine; T, thymine.
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
brane filters (Millipore Corp., Bedford, Mass.). The
filters were washed two times with a total volume of
0.7 ml of 5% (wt/vol) trichloroacetic acid, and the
radioactivity of the combined supernatant and washes
was measured. Control experiments showed that the
rate of release of acid-soluble radioactivity was linear
for at least 30 min. Activity is expressed as counts per
minute released per minute per milliliter of extract.
Chemical assays. DNA was determined by the
Burton diphenylamine method (8) with deoxyadenosine as the standard, using a factor of 2.4 to convert
deoxyadenosine values to DNA values. Protein was
determined by the method of Lowry et al. (23) with
bovine serum albumin as the standard.
Radioactivity. Radioactivity was determined by
scintillation counting. Samples from gradients, on
Whatmann 3MM filter disks, were counted in 7 ml of
toluene cocktail [toluene, 3.9 liters; 2,5-diphenyloxazole, 15.2 g; and 1,4-bis-2-(4-methyl-5-phenyl-axazolyl)-benzene, 1.4 g]. All other samples were counted in
5 ml of PCS (Amersham-Searle, Arlington Heights,
Ill.) solubilizer cocktail.
623
624
ROSSON AND RITTENBERG
J. BACTERIOL.
0
0
3
40~
31
>1
o2
0
0B
10
20
30
40
50
FRACTION NUMBER
FIG. 1. Sucrose density gradient profiles of [2'4C]deoxythymidine-labeled E. coli DNA taken at 60
min from a synchronous single-cycle B. bacteriovorus
culture. The initial culture contained 7.5 x 109 B.
bacteriovorus and 5 x 10i E. coli cells per ml (406,000
cpm). Samples (0.1 ml) were prepared as described in
the text and layered on neutral (6,600 cpm) or alkaline
(7,300 cpm) sucrose gradients. The gradients were
centrifuged at 20°C in a Beckman SW50.1 rotor at
50,000 rpm for 1.5 h (neutral gradient, 0) or 50,000
rpm for 2.33 h (alkaline gradient, 0). The positions of
the [methyl-3HJdeoxythymidine-labeled 029 reference DNA are indicated by arrows. The label B on
the abscissa indicates the centrifuge tube bottom for
this and other sedimentation profiles.
different experiments (R. A. Rosson, Ph.D. thesis, University of California, Los Angeles, 1978).
It was surprising that 60-min substrate cell
native DNA fragments of 5 x 105 daltons, representing a polynucleotide of approximately 780
base pairs, would not be bandable in CsCl, as
was previously reported by Matin and Rittenberg (25). Nevertheless, when 60-min DNA was
purified by their procedure it was not bandable,
nor could it be shown in any part of the gradient
over background. However, if a 60-min DNA
sample was prepared as described above, banded
on a 5 to 20% sucrose gradient, and sedimented
in CsCl to equilibrium, the DNA formed a diffuse band covering a large range in density,
typical of sheared double-stranded DNA (26).
The precise explanation for these different sedimentation characteristics of the degraded DNA
is unclear, but the difference was obviously related to the method of sample preparation. It is
clear, however, that the general conclusion was
correct: the substrate organism DNA is degraded to an intermediate size, with little concomitant production of soluble products.
Deoxyribonuclease activities of intraperiplasmically grown cells of B. bacteriovorus. Although it is known that axenically grown
cultures of B. bacteriovorus mutants excrete a
variety of hydrolytic enzymes, including nucleases, during growth (11, 12, 19), similar information is largely lacking for intraperiplasmicaily grown cells. Consequently, extracts of
bdellovibrios harvested just after the completion
of an intraperiplasmic growth cycle were assayed
for DNase activities. The assay procedures were
chosen on the assumption that the major bdellovibrio DNases would be similar to those found
in E. coli, the organism used as a control. The
measured activities of the controls (Table 2)
were in good agreement with published values
(35). The results of these assays showed that
intraperiplasmically grown bdellovibrios and
stationary-phase E. coli have similar levels of
exonuclease I, II, and III activities. In E. coli the
most active DNase is endonuclease I (22). The
activity of this enzyme in bdellovibrio extracts,
although less than 3% of the E. coli control level,
was clearly present. Although it is not suggested
that any of the enzymes assayed are necessarily
involved in the degradation of substrate cell
DNA, the data show that either preexisting E.
coli DNases or bdellovibrio-synthesized DNases
are potentially capable of degrading the substrate organism DNA.
Kinetics of degradation of E. coli DNA. A
synchronous bdellovibrio culture growing on
DNA-labeled E. coli was sampled at 15-min
intervals over the first 60 min of a 3-h cycle.
Portions removed from the same culture at 0,
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
precipitable fraction of degraded E. coli DNA
was established by rate zonal density gradient
centrifugation of DNA samples isolated from 60min bdellovibrio cultures. The neutral (native
DNA) and alkaline gradient profiles obtained
(Fig. 1) showed that the 60-min cold-trichloroacetic acid-insoluble degraded DNA sedimented
as a single, essentially homogeneous band under
both conditions. The estimated average molecular weights of the native and alkaline-denatured DNA were 3.0 x 106 and 1.4 x 106, respectively. The ratio of these molecular weights, 2:1,
indicates that the breakdown products were unnicked double-stranded fragments. The average
molecular weights of the 60-min fragments varied between 5 x 105 and 3.0 x 106 from experiment to experiment, apparently due to differences in the duration of the growth cycle in
DNA BREAKDOWN DURING B. BACTERIOVORUS GROWTH
VOL. 140, 1979
625
'
l
'60 J
-N-2
25-N-1
15 min
40 (20 K)
30-
50
0mi
20 (20 K)
15-
-j10
20-
N- , _. 8 N-6*^e.e
6
G6-45min
K(50
60min
(50K
Kl
4K)
26
6
-
d-3 4v-
A
I
\
FRACTION NUMBER
FIG. 2. Changes in native and alkaline-denatured E. coli DNA during intraperiplasmic growth of B.
bacteriovorus on [2- 4Cldeoxythymidine-labeled E. coli. The synchronous single-cycle culture originally
contained 1010 B. bacteriovorus and 5 x 10i E. coli cells per ml (100,000 cpm). Samples (0.2 ml) of the culture
were taken at the indicated times and lysed as described (see the text) with sodium dodecyl sulfate after
addition of [methyl- 3H]deoxythymidine-labeled T4D or 429 DNA. Either 0.1- or 0.2-ml portions (3,500 or 7,000
cpm) of the lysates were layered on neutral (N-1 to N-6) or alkaline (A-1 to A-6) sucrose gradients and
centrifuged in a Beckman SW50.1 rotor at 20°C. Neutral gradients were centrifuged at 20,000 rpm (20K) for
1.25 h or 50,000 rpm (50K) for 2.5 h; alkaline gradients were centrifuged at 20K or 35,000 rpm (35K) for 3 h or
50K for 2.5 h, as indicated. The positions of the [methyl-3H]deoxythymidine-labeled T4D (20K and 35K
gradients) or 4h29 (50K gradients) reference DNAs are indicated by arrows. Radioactivity of fractions is
plotted as percentage of total counts recovered.
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
15, 30, and 45 min were treated with chloram- banded in neutral and alkaline sucrose gradients.
phenicol and reincubated until 60 min after 0 The gradient profiles from the normal culture
time. DNA was isolated from all samples and are shown in Fig. 2, and those from the antibiotic-treated culture samples are shown in Fig.
3. The average molecular weights of the DNA
TABLE 2. DNase activities of stationary-phase E.
fragments calculated from these gradients and
coli and intraperiplasmically grown B.
other pertinent data are given in Tables 3 and 4.
bacteriovorus measured in crude cell-free extracts
With one exception (see Materials and MethSp acta
ods), essentially all radioactivity added was reExtract source
covered in both the neutral and alkaline graEndonuclease Exonuclease
dients (Tables 3 and 4). In all but the 15-min
I
I
III
II
alkaline and 30-min neutral gradients, most of
11 1
3
E. coli ML35
26
the radioactivity was found in sharply defined,
14 0.8 2.3
B. bacteriovorus
0.7
fairly symmetrical peaks, which suggested that
109J
DNA degradation was proceeding through disAs nanomoles of DNA nucleotide solubilized per crete molecular-sized fragments.
30 min per milligram of protein.
By 15 min in the untreated culture, the bdello-
626
ROSSON AND RITTENBERG
J. BACTERIOL.
------ii
~
"
||
/lr. -'I I
I\i 4[;iyLL
45in1hI1BLAlll
4 A-1
5pA-1N-58 N-64 1
..
A-2
o 8 -130min
45min
5 15min
i
0min
3
2 (50K)
.
21
(50K)
(20K)
0
Ho 10
0
14
J~~~~~~~~~~~~~~~~~~k
I
B.A530
3
2
2 A-
are
The hatched
indicated by
A4btpA-6
5
15m6-6
30me
45me
4 30mi
(35 K)
(50 K)
4(35 K) .A
50K)
4
f
oA.A3
Neura
.
(N- to fN-6J
and alau Al oA6MuroedniygrdetpoilefV
. 2
I
I
I
I
Cdoxt
.I
B 10 20 30 40 508 10'20 3040 SO B 10 20 30 40 508 10 20 30 40 50
FRACTION NUMBER
FIG. 3. Neutral (N-I to N-6) and alkaline (A-i to A-6) sucrose density gradient profiles of[2- 14CJdeoxythymidine-labeled E. coli DNA taken at 60min from a synchronous single-cycle B. bacteriovorus culture treated
with chloramphenicol at the times shown. Except for antibiotic addition, the experimental conditions were as
described in Fig. 2. The positions9 of the T4D (20K and 35K gradients) or q029 (50K gradients) reference DNAs
are indicated by arrows. The hatched bar8 show the center position of E. coli DNA (50K gradients) fr-om the
60-min untreated culture.
TABLE 3. Average molecular weight and native to alkaline-denatured DNA molecular weight ratios of
degraded E. coli DNA during intraperiplasmic growth of B. bacteriovorus"
Neutral gradients
Sample time
Alkaline gradients
Radioactivity
Radioactivity
recovered'
Avg mol wt
0
100
15
103
4.9 X 10
4.9 x 10W
x 10
x 10"
x 10"
x 10W
Radioactivity
vedy
rc
73
96
77
115
Radioactivity
50
29
109
110
21
20
79
100
N/A"
Avg mol wt
under peak
2.4 X 10'
2.9 x 10'
8.9 X 10'
5.2 x 10'
8.6 x 10"
5.4 x I0"
3.3 x I0"
70
2.0
42
64
50
11
5.5
25.0
79
74
1.7
1.7
74
34
121
45
60
93
from the neutral and alkaline gradient profiles of Fig. 2.
"Calculated
'
The ratio of average molecular weights of native to alkaline-degraded DNA.
'As percentage of total radioactivity in sample layered on gradient.
30
1.3
1.3
9.4
5.6
under peak"
1.7
1.5
vibrios had attached, penetrated, and converted change was detected in the sedimentation rate
the E. coli cells to the spherical form of the of native E. coli DNA (Fig. 2, N-1 and N-2).
bdelloplast. Over this period, no significant This did not hold foralkaline-dissociated DNA.
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
4
> -~
I-
VOL. 140, 1979
DNA BREAKDOWN DURING B. BACTERIOVORUS GROWTH
DNA fragments had been reduced by some 50%
to 5 x 105.
As a comparison of the data in Fig. 2 and 3
and Tables 3 and 4 shows, the addition of chloramphenicol to a culture at any time between 0
and 45 min largely prevented subsequent
changes in the molecular weight of substrate cell
DNA. Neutral and alkaline DNA profiles, generated from cultures receiving the antibiotic at
0, 15, 30, and 45 min and further incubated to 60
min, resembled more closely the profiles from
control (untreated) cultures taken at the same
times than they did profiles of control cultures
sampled 15 min or more later. For example,
single-strand nicks were not detected in 60-min
samples of cultures treated with chloramphenicol at 0 time (Fig. 3, A-1; Table 4). The alkaline
profile obtained from the antibiotic-containing
sample was essentially the same as that of the 0time control (Fig. 2, A-1), and the N/A ratio was
close to 2. Likewise, although the neutral profiles
from cultures treated with chloramphenicol at
15 min (Fig. 3, N-2) showed the fonnation over
the next 45 min of some native fragments similar
in size to the larger fragments from an untreated
30-min culture, at least half of the native DNA
remained undegraded and the 106-dalton fragments did not appear. Similarly, chloramphenicol added at 45 min appeared to eliminate formation of the 5 x 105-dalton pieces seen in
uninhibited cultures at 60 min.
These data suggested that, for complete degradation of the E. coli DNA, DNase synthesis is
required over the entire period of DNA breakdown and that possibly two endonucleases, not
functional during the initial endonucleolytic
nicking of the DNA, were synthesized after 15
min. One activity was apparently expressed between 15 and 30 min, and the second was expressed between 45 and 60 min. We could obtain
no evidence of additional endonucleolytic cuts
TABLE 4. Average molecular weights and native to alkaline-denatured DNA molecular uweight ratios for
degraded E. coli DNA from 60-min intraperiplasmic B. bacteriovorus cultures treated uith chloramphenicol
at indicated times"
Neutral gradients
CAM added
at (min):
0
15
Avg mol wt
5.3 x 10
4.7 x 10
5.0 x _0'-10 X 10
5.1 x 10
8.1 x 10"
8.9 x 10"
Alkaline gradients
Radioactivity re-
covered'"C,
113
Avg mol wt
2.9 x 10"
Radioactivity
covered' ('Zr)
98
78
9.0 x 10'-30 x 10'
1.0 X 107
30
70
4.0 X 10"
65
63
8.5 x 10"
70
45
102
4.6 x 10"
95
"Calculated
from neutral and alkaline gradient profiles of Fig. 3. CAM, Chloramphenicol.
'
The ratio of native to alkaline-denatured DNA molecular weight.
' As percentage of radioactivity in sample layered on gradient.
103
N/A
1.8
4-5
5-10
13
1.0
1.9
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
The 15-min alkaline profile, in contrast to the
corresponding 0-time profile, was asymmetric
(Fig. 2, A-1 and A-2). Most of the DNA (ca. 60%)
banded in the upper portion of the gradient and
had an average molecular weight significantly
less than the starting alkaline-dissociated material (Table 3). There remained, however, a considerable amount of DNA, centered in the middle of the gradient, which sedimented at the 0time rate. Comparison of the native to alkalinedissociated DNA (N/A) molecular weight ratios
for 0- and 15-min bandable DNA fractions (Table 3) gave quantitative evidence that the 15min native DNA had accumulated a large number of single-strand nicks not present at 0 time.
The results indicated an initial endonucleolytic
attack on substrate cell DNA.
Between 15 and 60 min into the growth cycle
of the bdellovibrios, the DNA of the substrate
cells was rapidly degraded. By 30 min, two major
size classes of native DNA of average molecular
weights of about 1.3 x 108 and 1.3 x 106 had
accumulated (Fig. 3, N-3 and N-4; Table 3).
Additionally, about half of the substrate cell
DNA consisted of higher-molecular-weight fragments more or less evenly distributed over the
bottom half of the neutral gradient (Fig. 2, N-3).
Analysis of the alkaline gradient profile (Fig. 3,
A-3 and A-4; Table 3) indicated that the highermolecular-weight fragments, 1.3 x 108 and
greater, contained multiple single-strand nicks
(N/A = 25), whereas the lower-molecularweight fraction was unnicked (N/A = 1.5). By
45 min, only a single peak of native DNA fragments was observed. This material comprised
about 80% of the total radioactivity and had
about the same average molecular weight and
N/A ratio as the small fragments at 30 min. The
results obtained with the 60-min sample were
very similar to those from the 45-min sample
except that the average molecular weight of the
627
628
ROSSON AND RITTENBERG
synthesis.
Figure 5 shows the rate of release of coldtrichloroacetic acid-soluble DNA fragments during bdellovibrio growth on heat-treated cells.
The experimental results shown are directly
comparable with those shown in Fig. 4 since the
same cell suspensions were used in both. The
rate of release of acid-soluble products started
earlier (15 min versus 30 min), and the rate and
total quantity of these products released were
always greater with the heated cells than with
unheated substrate cells. These differences were
probably due to single-strand nicks introduced
into the substrate organism's DNA as a result of
the heat treatment, which makes the E. coli
DNA a better substrate for exonuclease digestion (33). Conceivably, the presence of prenicked
DNA in the substrate cell also triggered the
production of exonucleases earlier than normal.
In any case, as in the experiments using unheated substrate cells, the addition of antibiotic
at 15 min or later prevented a further increase
in rate and caused the rate to drop rapidly.
These experiments confirmed that the bulk of
the exonuclease activity occurred after 30 min
with unheated substrate cells. The data also
indicated that continuous synthesis of exonucleases is necessary for solubilization of DNA of
both unheated and heated substrate cells, as was
also true for the synthesis of endonucleases involved in the degradation of high-molecularweight DNA. Since unheated substrate cells are
rapidly rendered incapable of protein synthesis
by bdellovibrio attack (39), and since heated
cells have neither biosynthetic potential nor appreciable levels of deoxynucleases, these data
suggest that both the endo- and exodeoxyribonucleases involved in DNA degradation are
bdellovibrio enzymes.
Total DNase activity during bdellovibrio
growth. As noted, the data shown in Fig. 4 and
5 indicate the balance between breakdown of
substrate cell DNA and DNA synthesis by the
bdellovibrios. To obtain a measure of the total
potential DNase activity of the bdellovibrio cultures, crude cell-free extracts were prepared
from cultures growing on untreated or heated E.
coli celLs and from the separate cell suspensions
used to start these cultures. The extracts were
supplemented with purified E. coli [3H]DNA,
and the rate of generation of cold-trichloroacetic
acid-soluble radioactivity was measured (Table
5). The sum of the individual activities of the E.
coli and bdellovibrio extracts, of the activity of
mixed extracts, and of the extract of the 0-time
culture were in good agreement, indicating that
both E. coli and B. bacteriovorus DNases can
be measured independently or together.
The potential DNase activity of bdellovibrio
cultures grown on heat-treated E. coli increased
slowly from 0 time on (Table 5). Treatment of
portions of the culture with chloramphenicol,
puromycin, or rifampin blocked further genera-
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
of the 5 x 105-dalton pieces, although such scissions could occur. Beyond 60 to 90 min, the
picture becomes obscured by the incorporation
of the 14C label into newly synthesized B. bacteriovorus DNA (25).
The rate of release of cold-trichloroacetic
acid-soluble DNA fragments from normal
and heated E. coli cells. The formation of
acid-soluble material from DNA is primarily a
function of exonuclease activity. During intraperiplasmic growth of the bdellovibrio, the rate
of release of cold-trichloroacetic acid-soluble radioactive material from DNA measures the difference between the rate of formation of soluble
fragments from substrate cell DNA and the rate
of their assimilation by the growing bdellovibrio.
As a second approach to identifying the origin
of the degradative DNases active during intraperiplasmic growth, a comparison was made of
release of cold-acid-soluble DNA fragments
from normal and heated E. coli and the effects
of chloramphenicol on this activity. Exploratory
experiments showed that essentially normal development patterns were obtained when bdellovibrios were grown on stationary-phase E. coli
cells preheated at 540C for 30 min (Rosson,
thesis; see also reference 14). The heated E. coli
were nonviable and retained little endonuclease
or exonuclease activity. Crude cell-free extracts
of the heated cells had less than 3% of the control
rates for endonuclease I and exonuclease activities.
The initial rate of release of cold-trichloroacetic acid-soluble DNA material from normal E.
coli cells is very low over the first 30 min of a
normal synchronous single-cycle culture. In the
experiment shown (Fig. 4), the rate amounted to
less than 0.03% of the total E. coli DNA initially
present, per min. After about 30 min the rate of
release increased sharply and reached a plateau
value some five times the initial rate. Concomitantly with the onset of bdelloplast lysis and
continuing until lysis was complete, the rate of
release again increased markedly. With the exhaustion of substrate at the end of the growth
cycle, the rate rapidly dropped to zero.
Chloramphenicol addition to bdellovibrio cultures at any time until lysis was well under way
not only prevented further increase in the rate
of release of cold-trichloroacetic acid-soluble
DNA products, but also caused an immediate
reduction in this rate. Essentially identical results were obtained with a second protein synthesis inhibitor, puromycin at 500 ,tg/ml, and
with rifampin, 100 ,ug/ml, an inhibitor of RNA
J. BACTERIOL.
DNA BREAKDOWN DURING B. BACTERIOVORUS GROWTH
VOL. 140, 1979
629
z
N..
I
O
.
z
Cz
'I
0
-J
0
0
cnol
0
0
-J
lJI
OL
LU
LLc
0
5t0)
TIME (min)
FIG. 4. Rate of release of cold-acid-soluble E. coli DNA products with time during synchronous growth of
B. bacteriovorus on [2- 14C]deoxythymidine-labeled E. coli in the presence or absence of chloramphenicol. The
initial culture contained 1010 B. bacteriovorus and 5 x 109 E. coli cells per ml (184,000 cpm). Portions of the
culture were treated with chloramphenicol (100 p g/ml) at various times. Samples of the cultures were mixed
with ice-cold trichloroacetic acid at the indicated times, and radioactivity was determined. Lysis was apparent
by 105 min. The arrows and symbols indicate the times of chloramphenicol addition: 0, untreated; 0, 0 min;
*, 15 min; A, 30 min; E, 60 min; V, 120 min.
z
N.
I-
-J
0
z
0-
z
0-
W
0
LC
CS)
.
a] (cL
mO
iiJ
-LJ
cJ 1
COWc
QO
LU
LU
(
0
LU
0
0
I-
5y
TIME (min)
FIG. 5. Rate of release of cold-acid-soluble E. coli DNA products with time during synchronous growth of
B. bacteriovorus on heat-treated, [2- 4CJdeoxythymidine-labeled E. coli in the presence or absence of
chloramphenicol. Experimental conditions and symbols are as in Fig. 4 except that the E. coli cells were
preheated at 54°C for 20 min.
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
LUJ
O
DoUf)
I I
630
ROSSON AND RITTENBERG
J. BACTERIOL.
TABLE 5. DNase activity in cell-free extracts of B.
bacteriovorus cultures growing intraperiplasmically
on unheated or heated E. colia
DNase activityb
Extract source
E. coli E. coli
heated heated
E. coli suspension
B. bacteriovorus suspension
Mixed extracts from no. 1 and 2
B. bacteriovorus culture, 0 min
B. bacteriovorus culture, 30 min
B. bacteriovorus culture, 45 min
B. bacteriovorus culture, 60 min
B. bacteriovorus culture, 90 min
As no. 8 with Cam at 30 min
As no. 8 with Pur at 30 min
As no. 8 with Rif at 30 min
6
72
96
104
153
253
373
452
116
101
136
626
220
862
905
530
377
652
910
361
361
397
a
Extracts were supplemented with [3H]DNA (4.6
x I04 cpm/,ug, unheated, or 9.0 x 103 cpm/pg, heated),
and the rate of 3H solubilization was measured. Anti-
biotic concentrations were: chloramphenicol (Cam),
200 ug/ml; puromycin (Pur), 500 ag/ml; and rifampin
100 Ag/ml.
(Rif),
I
Counts per minute solubilized per minute per milliliter of extract; average of two parallel, independent
samples.
tion of DNase activity. Relative to the rate of
activity measured at the time of antibiotic treatment, rates of activities in treated cultures declined by 10% (rifampin treated) to 34% (puromycin treated) after inhibitor addition. Control
experiments showed that the antibiotics used
had no direct effect on the measured rates when
added to extracts directly or to the cell suspensions immediately before extract preparation.
The DNase activity of extracts of bdellovibrios grown on unheated E. coli showed a
different pattern of change (Table 5). The rate
of generation of acid-soluble radioactivity by
samples taken between 0 and 45 min markedly
declined and then increased rapidly after 45 min.
The decline occurred over a period in which
synthesis of DNases by the intraperiplasmically
growing bdellovibrio and an increase in their
activity occurred (see Fig. 2 and 3). Since the
minimum value had a measured rate much less
than that of the initial E. coli (Table 4), the data
indicate that E. coli enzymes were rapidly inactivated over the first 45 min. Addition of inhibitors to these cultures had the same effects
noted with heated substrate cells: blocking of
further generation of DNase activity and a decline in activity existing at the time of addition.
DISCUSSION
The data presented permit two general conclusions: first, the regulated breakdown of sub-
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
strate cell DNA during intraperiplasmic growth
of B. bacteriovorus 109J is a consequence of the
sequential synthesis of several DNases of differing specificities by the bdellovibrio; and second,
preexisting DNases of the substrate cell are not
involved in any significant way in the degradative processes.
The first conclusion rests primarily on the
kinetic pattem of DNA breakdown and on the
effect of chloramphenicol on this pattern. The
sequential occurrence of discrete molecular
weight classes of fragments, the change in ratio
of the molecular weights of native to alkalinedegraded DNA from about 2 to 25 and back to
about 2, and the inhibition of further breakdown
of DNA by the addition of chloramphenicol to
a culture, not only support the conclusion but
also define qualitatively the activites involved.
The large increase in the N/A ratio in the absence of any detectable change in the average
molecular weight of native DNA over the first
15 min of the development cycle indicates that
the initial activity is an endonucleolytic singlestrand nicking of the DNA. The rapid buildup
of single-strand nicks and then the sudden emergence at about 30 min of an apparently unnicked
lower-molecular-weight fraction implies an endonuclease acting on the second strand opposite
or nearly opposite the initial nicks. Since chloramphenicol addition at 15 min permits a further
increase in the N/A ratio but no reappearance
of unnicked DNA fragments upon continued
incubation of the treated culture, it seems reasonable that this activity results from a new
endonuclease of altered specificity which is synthesized between 15 and 30 min. Similar arguments suggest that an additional endonuclease
that produces double-strand cuts may be synthesized between 45 and 60 min. Thus three
different sequential activities representing different enzymes could be involved in the initial
breakdown of the substrate organism's DNA.
The final products of these activities are fragments with average molecular weights of about
5 x 105. No fragments between this molecular
weight and acid-soluble size were detected,
which suggests that further attack involves only
exonuclease activity.
Exonuclease activity as measured by release
of acid-soluble DNA products first reaches appreciable levels between 30 and 45 min. This is
about as early as DNA synthesis can be detected
in the growing bdellovibrio (25). Addition of
protein synthesis inhibitors as late as 15 min
after initiation of a synchronous culture blocks
the release of acid-soluble DNA material, which
argues for a newly synthesized exonuclease and
not a preexisting one. Continued synthesis of
exonuclease apparently continues throughout
the total growth phase, since chloramphenicol
VOL. 140, 1979
DNA BREAKDOWN DURING B. BACTERIOVORUS GROWTH
Table 5). Under the conditions of the experiment, about 75% of the DNase activity in extracts of zero-time cultures is contributed by the
E. coli cells in the culture. By 45 min into the
development cycle, activities of culture extracts
have declined by about 60%. It is apparent from
quantitative considerations alone that most, if
not all, of the decrease must be due to inactivation of the E. coli enzymes. The quantitive argument is reinforced by the observation that
over the same period, bdellovibrios growing on
normal (Fig. 2 and 3) or heated (Table 5) cells
are synthesizing DNases.
A possible exception to the conclusion that
substrate cell enzymes do not function in DNA
degradation relates to the initial activity responsible for the multiple single-strand nicking. Although chloramphenicol added to a culture at
zero time eliminates this activity, the addition
also prevents penetration of the bdellovibrio into
the substrate cell. One can argue that penetration triggers endonuclease activity by disrupting
the normal control mechanisms of the substrate
cell, and therefore the effect of chloramphenicol
is indirect.
It will be noted that addition of chloramphenicol to a culture not only prevents the further
progression of the DNA degradative processes
but also results in a very marked decrease in the
rate of the ongoing degradative step. This was
clearly seen in the measurements of the rate of
release of cold-acid-soluble products from E. coli
DNA during bdellovibrio growth (Fig. 4). Chloramphenicol addition at any time until late in the
development cycle both prevented an increase
in the rate of this activity and also caused a very
rapid and substantial decrease in the rate. Addition of chloramphenicol to single-cycle cultures has a very similar effect on the solubilization of bdelloplast peptidoglycan during the lysis
stage of intraperiplasmic development (38). The
reasons for the rapid decrease in these activities
is not completely clear, but at least two factors
are probably involved. One relates to the nature
of a single-cycle bdellovibrio culture, which is
not a single culture but rather some 109 individual cultures isolated from one another. Since
each bdellovibrio is growing within a chamber
from which macromolecules apparently do not
escape until lysis (15, 25, 29), an enzyme functioning in a particular bdelloplast does not have
access to substrate in a second. Therefore, addition of chloramphenicol to a culture at a particular time would not only prevent further enzyme synthesis but would also limit continuing
activity to those bdelloplasts in which the measured activity had been initiated. Thus rates
could decline well before total potential substrate in the overall culture was consumed. The
magnitude of this effect would depend on the
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
addition at any point up to bdelloplast lysis
prevents further release of acid-soluble DNA
material.
A similar pattern of degradation of DNA to a
discrete cold-acid-insoluble size, followed by exonuclease digestion, has been observed in a number of systems. In less than 15 min E. coli
bacteriophage T4 rapidly degrades the DNA of
its host cell to native fragments of 200 x 10'
daltons, and by 30 min DNA is degraded to
about 106 daltons (21, 38, 42). These fragments
are rapidly solubilized and immediately incorporated into phage progeny DNA. As in the
results reported here, no fragments between 106
daltons and acid-soluble size accumulate. The
majority of this degradation is the result of T4
DNase activity (5, 41). Similar degradation of
host cell DNA by phage-induced enzymes has
been observed in other bacteriophage systems
(9, 13, 32, 36). The antimicrobial agent colicin
E2 also induces the rapid degradation of E. coli
DNA to fragments of about 106 daltons (18, 29).
These DNA fragments are then solubilized to
acid-soluble form, also without any smaller fragments being observed. In the colicin system,
however, the E. coli enzymes are responsible for
the DNA degradation (2, 18, 29).
It will be noted that over the period between
60 and 120 min the rate of release of acid-soluble
DNA material by bdellovibrios growing on normal unheated cells is relatively constant and
amounts to about 0.14% of the total substrate
cell DNA per min. During the same time interval, over 50% of the substrate cell DNA is incorporated into bdellovibrio DNA (25), i.e., a rate
of about 0.85% per min of the total available.
Assuming all DNA precursors are incorporated
as nucleoside monophosphates (30), the actual
rate of exonuclease activity is about seven times
that of the measured release of the cold-acidsoluble material over this period of the development cycle, and the efficiency of assimilation
is about 85%.
Since bdellovibrios grow with normal kinetics,
efficiency, and cell yields on properly heated
substrate cells (14; Rosson, thesis) in which essentially all deoxynuclease activity has been
eliminated, it is clear that these substrate cell
enzymes are not essential for intraperiplasmic
growth. Further, accepting that the sequential
synthesis of several DNases is required for DNA
breakdown, it follows that substrate cell enzymes are not functioning in the process even
when the bdellovibrio is growing on normal cells,
since the attacked cell is rapidly rendered incapable of energy generation (31) and protein synthesis (39). Furthermore, our data show directly
that preexisting deoxynucleases of unheated
substrate cells are inactivated over the first 45
min of the intraperiplasmic growth cycle (see
631
632
ROSSON AND RITTENBERG
ACKNOWLEDGMENTIS
This work was supported by grant PCM75-18883 from the
National Science Foundation. R.A.R. held a Public Health
Service Postdoctoral Traineeship (GM-01297) from the National Institute of General Medical Sciences during part of the
period of this investigation.
LITERATURE CITED
1. Adams, M. H. 1959. Bacteriophages. Interscience Publishers, Inc., New York.
2. Almendinger, R., and L P. Hager. 1973. Reconstitution
of colicin E2-induced deoxyribonucleic acid degradation
in spheroplast preparations. Antimicrob. Agents Chemother. 4:167-177.
3. Anderson, D. L, and E. T. Mosharrafa. 1968. Physical
and biological properties of phage 29 deoxyribonucleic
acid. J. Virol. 2:1185-1190.
4. Bollum, F. J. 1968. Filter paper disc techniques for assaying radioactive macromolecules. Methods Enzymol.
12B:169-173.
5. Bose, S. K., and R. J. Warren. 1969. Bacteriophageinduced inhibition of host functions. H. Evidence for
multiple, sequential bacteriophage-induced deoxyribonucleases responsible for degradation of cellular deoxyribonucleic acid. J. Virol. 3:549-566.
6. Boyce, R. P., and R. B. Setlow. 1962. A simple method
of increasing the incorporation of thymidine into the
deoxyribonucleic acid of Escherichia coli. Biochim.
Biophys. Acta 61:618-620.
7. Burgi, E., and A. D. Hershey. 1963. Sedimentation rate
as a measure of molecular weight of DNA. Biophys. J.
3:309-321.
8. Burton, K. 1956. A study of the conditions and mechanism of the diphenylamine reaction for the colorimetric
estimation of deoxyribonucleic acid. Biochem. J. 62:
315-323.
9. Crawford, L V. 1959. Nucleic acid metabolism in Escherichia coli infected with phage T5. Virology 7:359-374.
10. Eigner, J., and P. Doty. 1965. The native, denatured
and renatured states of deoxyribonucleic acid. J. Mol.
Biol. 12:549-580.
11. Engelking, H. M., and R. Seidler. 1973. The involvement of extracellular enzymes in the metabolism of
Bdellovibrio. Arch. Microbiol. 95:293-304.
12. Gloor, L, B. Klubeck, and R. J. Seidler. 1973. Molecular heterogeneity of the Bdellovibrios: metallo and
serine proteases unique to each species. Arch. Microbiol. 95:45-56.
13. Hausmann, R., and B. Gomez. 1968. Bacteriophage T3and T7-directed deoxyribonucleases. J. Virol. 2:265266.
14. Hespell, R. B. 1978. Intraperiplasmic growth of Bdellovibrio bacteriovorus on heat-treated Escherichia coli.
J. Bacteriol. 133:1156-1162.
15. Hespell, R. B., G. F. Miozzari, and S. C. Rittenberg.
1975. Ribonucleic acid destruction and synthesis duTing
intraperiplasmic growth of Bdellovibrio bacteriovorus.
J. Bacteriol. 123:481491.
16. Hespell, R. B., and D. A. Odelson. 1978. Metabolism of
RNA-ribose by Bdellovibrio bacteriovorus during intraperiplasmic growth on Escherichia coli. J. Bacteriol.
136:936-946.
17. Hespell, R. B., R. A. Rosson, M. F. Thomashow, and
S. C. Rittenberg. 1973. Respiration of Bdellovibrio
bacteriovorus strain 109J and its energy substrates for
intraperiplasmic growth. J. Bacteriol. 113:1280-1288.
18. Holland, E. M., and I. B. Holland. 1972. Kinetics of
colicin E2-induced solubiliation and fragmentation of
Escherichia coli DNA in vivo. Biochim. Biophys. Acta
281:179-191.
19. Huang, J. C. C., and M. P. Starr. 1973. Possible enzymatic bases of bacteriolysis by bdellovibrios. Arch. Mikrobiol. 89: 147-167.
20. Jacob, A. E., and S. J. Hobbs. 1974. Conjugal transfer
of plasmid-borne multiple antibiotic resistance in Streptococcus faecalis var. zymogenes. J. Bacteriol. 117:360372.
21. Kutter, E. M., and J. S. Wiberg. 1968. Degradation of
cytosine-containing bacterial and bacteriophage DNA
after infection of Escherichia coli B with bacteriophage
T4D wild type and with mutants defective in genes 46,
47, and 56. J. Mol. Biol. 38:395411.
22. Lehman, L. R., G. G. Rousses, and E. A. Pratt. 1962.
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
precise time in the cycle the inhibitor was added
and the degree of synchrony of the culture.
A second and probably much more important
reason also relates to the bdellovibrio's growth
in an isolated chamber. The exoenzymes responsible for the degradation of substrate cell macromolecules must be released by the bdellovibrio
into what was the periplasmic space or the protoplast of the attacked cell or into both and must
be retained there. Included among these enzymes would be proteases, known to be produced
by the bdellovibrios (11, 12, 19), which generate
the amino acids used for biosynthesis and energy
production (17). These could act not only on
substrate cell proteins, including the substrate
cell's DNases, whose inactivation was discussed
above, but also on the other exoenzymes introduced by the bdellovibrio into the bdelloplast
chamber. As a consequence, continuous production of these enzymes would be required to sustain the degradative processes, and inhibition of
their synthesis could result in the observed rapid
decrease in the rate of these processes.
In those experiments in which the total potential DNase activity was assayed using extracts
of cultures, the addition of chloramphenicol did
not result in as great a decline in rate (Table 5)
as the decline in rate of release of cold-acidsoluble DNA fragments when the antibiotic effect was measured with intact bdelloplasts (Fig.
4). Although the two sets of data may appear
contradictory, they are not. In the former case,
the sonic treatment of the bdelloplasts used to
prepare the crude extracts destroyed their normal compartmentalization, releasing intracellular enzymes of the developing bdellovibrios.
These enzymes would not be involved in the
degradative processes occurring in the bdelloplast chamber nor would they be susceptible to
digestion by the exoproteases released into the
chamber. Based on this analysis, the data in
Table 5 suggest that by 30 min, essentially all of
the original E. coli deoxynucleases, whether initially located periplasmically or intracellularly,
were destroyed or were susceptible to degradation, and that some 10 to 30% of the bdellovibrio
deoxynucleases had been released into the bdelloplast chamber.
J. BACTERIOL.
VOL. 140, 1979
23.
24.
25.
26.
28.
29.
30.
31.
32.
The deoxyribonucleases of Escherichia coli. II. Purification and properties of a ribonucleic acid-inhibitable
endonuclease. J. Biol. Chem. 237:819-828.
Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.
Randall. 1951. Protein measurement with the Folin
phenol reagent. J. Biol. Chem. 193:265-275.
Marmur, J. 1963. A procedure for the isolation of deoxyribonucleic acid from microorganisms. Methods Enzymol. 6:726-738.
Matin, A., and S. C. Rittenberg. 1972. Kinetics of deoxyribonucleic acid destruction and synthesis during
growth of Bdellovibrio bacteriovorus strain 109D on
Pseudomonasputida and Escherichia coli. J. Bacteriol.
111:664-673.
Meselson, M., F. W. Stahl, and J. Vinograd. 1957.
Equilibrium sedimentation of macromolecules in density gradients. Proc. Natl. Acad. Sci. U.S.A. 43:581-588.
Neu, H. 1967. The 5'-nucleotidase of Escherichia coli. I.
Purification and properties. J. Biol. Chem. 242:38963904.
Reilly, B. E., and J. Spizizen. 1965. Bacteriophage deoxyribonucleate infection of competent Bacillus subtilis.
J. Bacteriol. 89:782-790.
Ringrose, P. 1970. Sedimentation analysis of DNA degradation products resulting from the action of colicin
E2 on Escherichia coli. Biochim. Biophys. Acta 213:
320-334.
Rittenberg, S. C., and D. Langley. 1975. Utilization of
nucleoside monophosphates per se for intraperiplasmic
growth ofBdellovibrio bacteriovorus. J. Bacteriol. 121:
1137-1144.
Rittenberg, S. C., and M. Shilo. 1970. Early host damage
in the infection cycle of Bdellovibrio bacteriovorus. J.
Bacteriol. 102:149-160.
Sadowski, P. D., and C. Kerr. 1970. Degradation of
Escherichia coli B deoxyribonucleic acid after infection
with deoxyribonucleic acid-defective amber mutants of
633
bacteriophage T7. J. Virol. 6:149-155.
33. Sedgwick, S. G., and B. A. Bridges. 1972. Evidence for
indirect production of DNA strand scissions during mild
heating of Escherichia coli. J. Gen. Microbiol. 71:191193.
34. Shilo, M., and B. Bruff. 1965. Lysis of gram-negative
bacteria by host-independent ectoparasitic Bdellovibrio
bacteriovorus isolates. J. Gen. Microbiol. 40:317-328.
35. Shortman, K., and L. R. Lehman. 1964. The deoxyribonucleases of Escherichia coli. VI. Changes in enzyme
levels in response to alterations in physiological state.
J. Biol. Chem. 239:2964-2974.
36. Stone, A. B., and K. Burton. 1962. Studies on the
deoxyribonucleases of bacteriophage-infected Escherichia coli. Biochem. J. 85:600-606.
37. Studier, F. W. 1965. Sedimentation studies of the size
and shape of DNA. J. Mol. Biol. 11:373-390.
38. Thomashow, M. F., and S. C. Rittenberg. 1978. Intraperiplasmic growth of Bdellovibrio bacteriovorus 109J:
solubilization of Escherichia coli peptidoglycan. J. Bacteriol. 135:998-1007.
39. Varon, M., L. Drucker, and M. Shilo. 1969. Early effects
of Bdellovibrio infection on the synthesis of protein and
RNA of host bacteria. Biochem. Biophys. Res. Commun. 37:518-525.
40. Varon, M., and M. Shilo. 1968. Interaction of Bdellovibrio bacteriovorus and host bacteria. I. Kinetic studies of attachment and invasion of Escherichia coli B by
Bdellovibrio bacteriovorus. J. Bacteriol. 95:744-753.
41. Warner, H. R., D. P. Snustad, S. E. Jorgensen, and J.
F. Koerner. 1970. Isolation of bacteriophage T4 mutants defective in the ability to degrade host deoxyribonucleic acid. J. Virol. 5:700-708.
42. Warren, R. J., and S. K. Bose. 1968. Bacteriophageinduced inhibition of host functions. I. Degradation of
Escherichia coli deoxyribonucleic acid after T4 infection. J. Virol. 2:327-334.
Downloaded from http://jb.asm.org/ on February 6, 2015 by guest
27.
DNA BREAKDOWN DURING B. BACTERIOVORUS GROWTH