Response of Legionella pneumophila to β-lactam antibiotics

Abstract

Legionella pneumophila Philadelphia strain 1 grown in vitro contained five penicillin-binding proteins that were accessible
to the antibiotic in membrane preparations and in live cells as well. The bacterium had reasonably low MICs of several beta-lactam
antibiotics and was susceptible to both the bactericidal and the lytic activity of these drugs. An unusual feature of the
response of this bacterium to penicillin treatment was that cell lysis as determined by decrease in culture turbidity and
release of intracellular macromolecules was not accompanied by degradation of the peptidoglycan.

Full text

ANTIMICROBIAL
AGENTS
AND
CHEMOTHERAPY,
May
1985,
p.
695-700
Vol.
27,
No.
5
0066-4804/85/050695-06$02.00/0
Copyright
©
1985,
American
Society
for
Microbiology
Response
of
Legionella
pneumophila
to
,B-Lactam
Antibiotics
STEVEN
WEISHOLTZ
AND
ALEXANDER
TOMASZ*
The
Rockefeller
University,
New
York,
New
York
10021
Received
18
October
1984/Accepted
5
February
1985
LegioneUa
pneumophila
Philadelphia
strain
1
grown
in
vitro
contained
five
penicillin-binding
proteins
that
were
accessible
to
the
antibiotic
in
membrane
preparations
and
in
live
cells
as
well.
The
bacterium
had
reasonably
low
MICs
of
several
P-lactam
antibiotics
and
was
susceptible
to
both
the
bactericidal
and
the
lytic
activity
of
these
drugs.
An
unusual
feature
of
the
response
of
this
bacterium
to
penicillin
treatment
was
that
cell
lysis
as
determined
by
decrease
in
culture
turbidity
and
release
of
intracellular
macromolecules
was
not
accompanied
by
degradation
of
the
peptidoglycan.
Legionella
pneumophila,
although
similar
to
other
gram-
negative
bacteria
in
terms
of
its
basic
cell
envelope
struc-
tures
and
composition
(1,
8,
13),
has
several
unusual
char-
acteristics.
It
was
reported
that
cell
wall
preparations
from
L.
pneumophila
Philadelphia
strain
2
contain
strongly
bound
protein
material
that
is
not
removed
by
trypsin
but
only
by
alkali
treatment,
and
that
cleavage
of
the
glycan
chains
in
the
L.
pneumophila
wall
by
egg
white
lysozyme
does
not
cause
a
decrease
in
the
turbidity
of
cell
wall
suspensions
(1).
Because
an
unusual
cell
wall
structure
may
contribute
to
the
relative
inefficacy
of
P-lactam
antibiotics
against
these
path-
ogens
(2,
6,
9)
and
because
the
physiology
of
interactions
of
L.
pneumophila
with
these
antibiotics
has
not
been
well
documented,
we
undertook
a
morphological
and
physiolog-
ical
study
of
the
response
of
this
bacterium
to
some
,-lactam
antibiotics.
We
showed
that
L.
pneumophila
has
penicillin-
binding
proteins
(PBPs)
similar
to
those
found
in
other
gram-negative
bacilli.
The
bacteria
had
prompt
bactericidal
and
lytic
responses
to
various
,B-lactam
antibiotics.
How-
ever,
these
responses
were
unique
in
that
penicillin-induced
cell
lysis
was
not
accompanied
by
extensive
cell
wall
degra-
dation.
MATERIALS
AND
METHODS
Strains
and
media.
L.
pneumophila
Philadelphia
strain
1,
obtained
from
Marcus
Horwitz,
the
Rockefeller
University,
New
York,
N.Y.,
was
used
in
all
experiments.
The
organism
was
maintained
by
subculturing
onto
charcoal-yeast
extract
agar
(7)
every
72
h
after
growth
at
37°C
with
5%
CO2.
For
growth
experiments,
several
colonies
were
selected
from
the
agar
plate
and
inoculated
into
a
synthetic
medium
(15)
that
was
supplemented
with
yeast
extract (0.1%;
Difco
Labora-
tories,
Detroit,
Mich.),
bovine
serum
albumin
(0.4%),
L-Cys-
teine
(0.4
mg/ml),
and
ferric
pyrophosphate
(0.25
mg/ml).
This
medium,
to
be
referred
to
as
growth
medium
in
this
paper,
was
easy
to
prepare
and
supported
luxuriant
growth
of
L.
pneumophila
in
preliminary
experiments.
In
labeling
experiments
with
glucosamine,
the
basic
medium
was
used
except
that
all
glucose
was
omitted.
The
Escherichia
coli
strain
used
was
W7
(lys
dap).
Antibiotics
and
reagents.
Benzylpenicillin
G
(Bristol
Lab-
oratories,
Syracuse,
N.Y.),
aztreonam
(E.
R.
Squibb
&
Sons,
Princeton,
N.J.),
thienamycin,
mecillinam,
cefot-
axime,
and
cefoxitin
(Merck
&
Co.,
Inc.,
Rahway,
N.J.)
were
stored
as
powders
and
then
dissolved
in
water
at
the
time
of
an
experiment.
Other
chemicals,
reagents,
and
*
Corresponding
author.
medium
components
were
reagent-grade,
commercial
prod-
ucts.
MIC
assays.
All
MIC
assays
were
done
in
growth
medium
with
a
broth
macrodilution
method.
Twofold
serial
dilutions
of
the
antibiotics
were
made
in
broth
on
the
day
of
the
experiment.
The
organism
had
been
subcultured
from
a
charcoal-yeast
agar
plate
into
10
ml
of
growth
medium
and
grown
overnight
at
35°C
with
aeration.
A
1:10
dilution
of
this
culture
was
made
into
fresh,
prewarmed
medium
and
then
grown
back
over
a
period
of
4
to
5
h
to
log
phase.
Another
1:10
dilution
was
then
made,
and
0.1-ml
portions
of
this
diluted
culture
were
inoculated
into
the
antibiotic
tubes
containing
a
final
volume
of
1
ml.
The
final
inoculum
was
-105
CFU
per
MIC
tube.
Purity
of
the
cultures
was
ascer-
tained
by
microscopic
and
colonial
appearance
and
failure
to
grow
on
tryptic
soy
agar
and
blood
agar
plates.
The
MIC
tubes
were
incubated
at
37°C
with
5%
C02,
and
growth
was
assessed
at
48,
72,
and
96
h.
Samples
(100
ulI)
were
removed,
serially
diluted,
and
plated
for
the
determination
of
MBCs,
defined
as
the
lowest
antibiotic
concentration
that
gave
a
99.9%
kill
after
48
h
growth.
Colony
counts
were
done
after
plating
onto
charcoal-yeast
agar
and
incubating
at
37°C
with
5%
CO2
for
72
h.
Growth
and
lysis
experiments.
The
organism
was
subcul-
tured
from
an
agar
plate
into
10
ml
of
growth
medium
and
grown
overnight
at
37°C
with
aeration.
The
culture
was
then
backdiluted
into
fresh
medium
and
regrown
to
log
phase.
Growth
and
lysis
after
addition
of
antibiotic
was
monitored
by
using
a
Coleman
nephelometer
with
a
620-nm
filter.
One
hundred
nephelometric
turbidity
units
corresponded
to
-107
CFU/ml.
Cell
viability
was
determined
by
making
dilutions
in
growth
medium
and
then
plating
onto
charcoal-yeast
agar
plates.
Morphology.
To
assess
the
morphological
response
to
antibiotics,
1
ml
of
a
log-phase
culture
containing
-107
CFU/ml
was
dispensed
into
sterile
tubes,
and
antibiotics
were
added
in
concentrations
of
0.25,
0.50,
1,
and
1OX
MICs.
The
cultures
were
incubated
at
37°C
with
5%
CO2
for
24
h,
and
then
phase
microscopy
photographs
were
taken
with
a
Nikon
camera
and
Carl
Zeiss
microscope
at
X1,250
magnification.
Cultures
were
prepared
for
electron
micros-
copy
by
adding
glutaraldehyde
(2%)
and
by
chilling
on
ice
for
10
min.
The
cultures
were
then
filtered
over
a
45-,um
Millipore
membrane
filter
(Millipore
Corp.,
Bedford,
Mass.).
The
filter
was
placed
in
2%
glutaraldehyde
with
0.1
M
KPO4
buffer
(pH
7.0)
to
fix
the
cells.
Dehydration,
embedding,
and
thin
sectioning
were
done
as
described
before
(17).
Peptidoglycan
degradation
and
cell
lysis.
Log-phase
cul-
695

6%
WEISHOLTZ
AND
TOMASZ
tures
of
L.
pneumophila
were
labeled
by
growth
for
18
to
20
h
in
10-ml
batches
of
[3H]glucosamine
(19.7
Ci/mmol;
New
England
Nuclear
Corp.,
Boston,
Mass.)
at
2
p.Ci
and
2
p.g
of
medium
per
ml
and
no
glucose.
The
cells
were
then
centri-
fuged
at
7,500
rpm
at
4°C
and
suspended
in
50
ml
of
prewarmed
medium
to
which
50
p.g
of
nonradioactive
gluco-
samine
per
ml
had
been
added.
The
culture
was
then
incubated
for
4
to
5
h
to
ensure
utilization
of
intracellular
label.
Such
prelabeled,
exponentially
growing
cultures
re-
ceived
the
antibiotics
at
various
concentrations,
and
incuba-
tion
continued.
At
intervals,
1.0-mI
samples
were
removed
and
centrifuged
in
an
Eppendorf
microfuge
at
4°C
for
10
min.
The
supernatant
was
discarded,
and
the
pellet
was
sus-
pended
in
cold
H20,
transferred
to
borosilicate
tubes,
and
frozen.
After
thawing,
the
samples
were
boiled
in
4%
sodium
dodecyl
sulfate
(SDS)
for
30
min.
The
SDS-insoluble
mate-
rial
was
washed
three
times
with
H20
over
a
Millipore
filter
(0.45-p.m
pores).
The
filters
were
then
placed
in
vials
with
1
ml
of
potassium
phosphate
buffer
(50
mM,
pH
7.0)
contain-
ing
200
,ug
of
Ml
muramidase
and
were
gently
agitated
at
room
temperature
overnight.
This
procedure
resulted
in
the
complete
solubilization
of
radioactively
labeled
cell
walls.
Portions
of
the
solutions
were
assayed
for
radioactivity
in
15
ml
of
liquid
scintillation
fluid
(Ultrafluor;
National
Diagnos-
tics,
Somerville,
N.J.)
in
a
Nuclear-Chicago
Mark
II
scintil-
lation
counter.
For
RNA
labeling,
log-phase
cultures
were
labeled
during
overnight
growth
with
a
combination
of
[3H]uracil
(230
Ci/mmol;
New
England
Nuclear)
at
5
,ug
and
2.5
,uCi
of
medium
per
ml.
The
cells
were
centrifuged,
suspended
in
fresh
medium,
and
grown
for
4 h
to
allow
utilization
of
intracellular
label.
After
the
addition
of
antibiotic,
500-,ul
portions
were
removed
and
centrifuged
in
an
Eppendorf
microfuge
for
10
min
at
4°C.
Radioactivity
released
from
the
cells
during
antibiotic
treatment
was
determined.
A
250-,u
portion
of
the
supernatant
was
added
to
an
equal
volume
of
ice-cold
10%
trichloroacetic
acid
and
kept
at
4°C
overnight.
The
trichloroacetic
acid-precipitable
material
was
collected
on
Millipore
filters
(0.45-p.m
pores),
washed
three
times
with
H20,
and
dried
at
100°C
for
1
h.
A
toluene-based
fluor
was
added
to
the
filters,
and
radioactivity
was
determined
as
above.
Total
radioactivity
was
determined
by
trichloroacetic
acid
treatment
of
a
500-pul
sample
of
cells
before
antibiotic
exposure.
Preparation
of
murein
labeled
with
radioactive
glucos-
amine.
Cultures
were
prelabeled
with
[3H]glucosamine
as
above.
After
harvesting
by
centrifugation,
the
cells
were
suspended
in
H20
and
frozen.
After
thawing,
the
cells
were
boiled
in
4%
SDS
for
30
min
with
the
addition
of
50
mM
sodium
acetate
(pH
5.0)
to
minimize
loss
of
O-acetyl
groups
(3).
After
cooling,
the
SDS-insoluble
material
was
collected
by
ultracentrifugation
at
100,000
xg
for
30
min.
The
pellet
was
washed
with
H20
three
times
by
ultracentrifugation.
For
muramidase
digestion
assays,
100-pl
samples
of
purified
peptidoglycan
containing
-10,000
cpm
were
added
to
1
ml
of
buffer
and
various
enzymes
and
incubated
overnight.
The
following
enzymes
were
used:
trypsin
(in
100
mM
Tris,
pH
8),
crude
Ml
muramidase
(in
100
mM
potassium
phosphate,
pH
7),
purified
Ml
muramidase
(in
100
mM
potassium
phosphate,
pH
7),
or
egg
white
lysozyme
(in
50
mM
ammo-
nium
acetate,
pH
6.5)
at
100
p.g
per
ml
or
Chalaropsis
muramidase
(20
p.g/ml).
The
buffers
used
were
Tris,
KPO4,
and
sodium
acetate
(50
mM,
pH
5.0)
for
Chalaropsis
mur-
amidase.
The
digested
samples
were
boiled
in
4%
SDS
for
30
min.
The
SDS-insoluble
portion
was
then
removed
by
Mill-
ipore
filtration
(0.45-p.m
pores).
The
solubilized
material
was
TABLE
1.
MICs
and
MBCs
of
several
3-lactam
antibiotics
for
L.
pneumophila
Philadelphia
strain
1
Antibiotic
MIC48a
MIC72
MBC48
Benzylpenicillin
1.25
1.25
1.25
Aztreonam
5.0
5.0
10.0
Thienamycin
<0.02
0.02
-0.02
Mecillinam
1.25
10
>10
Cefotaxime
0.15
0.15
0.15
Cefoxitin
<0.02
0.15
0.3
a
MIC,",
MIC
after
48
h
of
exposure
to
antibiotic
as
described
in
the
text.
analyzed
by
thin-layer
chromatography
(TLC).
Undigested
murein
was
quantitated
by
determining
the
radioactivity
on
the
filters.
TLC.
The
procedure
of
Gmeiner
and
Kroll
was
used
(10).
Solubilized
peptidoglycan
(8,000
to
20,000
cpm)
was
applied
to
a
0.25-mm-thick
silica
gel-coated
plastic
sheet
(20
x
20
cm;
Eastern
Organic
Chemicals,
Rochester,
N.Y.).
The
TLC
plate
was
activated
before
sample
application
at
100°C
for
30
min.
The
chromatogram
was
developed
twice
in
isobutyric
acid-1
M
ammonia
(5:3,
vol/vol),
dried,
and
sprayed
with
En3Hance
(New
England
Nuclear),
and
the
fluorographs
were
exposed
for
6
days.
Radioactive
spots
were
cut
out,
placed
in
scintillation
vials,
and
counted
in
a
toluene-based
fluor.
PBPs.
After
overnight
growth,
the
culture
was
backdiluted
1:10
into
fresh
broth
and
grown
to
log
phase.
This
culture
was
concentrated
10-fold,
and
100-,u
volumes
were
exposed
to
various
concentrations
of
[3H]benzylpenicillin
(25
Ci/mmol;
Merck
&
Co.)
and
incubated
for
30
min
at
37°C.
The
cultures
were
then
chilled,
centrifuged
at
4°C
for
10
min
in
an
Eppendorf
microfuge,
and
washed
twice
with
1
ml
of
cold
potassium
phosphate
buffer
(50
mM,
pH
7.0)
to
remove
the
albumin
containing
medium.
The
sediment
was
sus-
pended
in
50
RI1
of
potassium
phosphate
buffer;
35
,ul
of
sample
buffer
was
added,
and
the
samples
were
boiled
for
5
min
(4).
They
were
subjected
to
polyacrylamide
gel
elec-
trophoresis
(PAGE)
for
the
detection
of
PBPs
as
previously
described
(16).
Fluorograms
were
exposed
for
12
days.
L.
pneumophila
membranes
were
prepared
from
50-ml
cultures
grown
overnight.
After
sedimentation
at
4°C,
the
cells
were
washed
three
times
in
20
ml
of
potassium
phos-
phate
buffer
(50
mM,
pH
7.0),
suspended
in
5
ml
of
H20
containing
4
mM
MgCI2,
and
frozen
at
-20°C.
After
thaw-
ing,
an
equal
volume
of
glass
beads
(100-p.m
diameter)
was
added,
and
the
mixture
was
shaken
on
a
Mickle
disintegrator
for
60
min
at
4°C.
DNase
and
RNase
(both
50
p.l
at
2
mg/ml)
were
added,
and
the
mixture
was
incubated
for
10
min
at
37°C.
The
suspension
was
then
centrifuged
at
5,000
x
g
for
10
min,
and
the
supernatant
was
collected
and
ultracentri-
fuged
at
40,000
rpm
for
40
min
at
4°C.
The
sediment
was
washed
three
times
in
H20,
suspended
in
1.2
ml
of
H20,
and
frozen
(1.3
mg
of
protein
per
ml).
For
the
PBP
assay,
260
p.g
of
membrane
was
incubated
with
various
concentrations
of
[3H]benzylpenicillin
for
20
min
at
37°C.
The
reaction
was
stopped
by
adding
10
p.l
of
Sarkosyl
and
5
p.1
of
nonradioactive
benzylpenicillin
(12.5
p.g/Il).
Next,
40
p.1
of
sample
buffer
was
added,
and
the
mixture
was
boiled
for
5
min.
The
sample
was
then
subjected
to
PAGE,
and
fluorograms
were
prepared
as
above.
For
the
determination
of
the
rates
of
deacylation
of
penicilloyl-PBPs,
the
membrane
preparation
was
incubated
for
20
min
with
a
saturating
concentration
of
[3H]benzylpen-
icillin
(20
,ug/ml)
followed
by
the
addition
of
nonradioactive
ANTIMICROB.
AGENTS
CHEMOTHER.

LYSIS
WITHOUT
WALL
DEGRADATION
IN
L.
PNEUMOPHILA
TABLE
2.
Ability
of
,-lactam
antibiotics
to
cause
lysis
of
L.
pneumophilaa
(h)
%
of
initial
turbidity
Time"
Penicillin
G
Thienamycin
Mecillinam
Aztreonam
Control
0
100 100
100
100
100
0.5
100
100 100
100
110
1
70 90 70
95
130
1.5
55
60
55
80
165
2
50
50
50
60
180
4
45
45
50
55
210
a
When
exponentially
growing
cultures,
with
shaking,
reached
a
cell
concentration
of
-5
x
10'
CFU/ml,
antibiotics
were
added,
each
at
a
concentration
of
lOx
MIC,
and
the
turbidity
was
monitored
with
a
Coleman
nephelometer.
b
Time
after
addition
of
antibiotic.
benzylpenicillin
(2.5
mg/ml),
and
incubatioh
at
37°C
contin-
ued.
Portions
(70-,ul)
were
removed
after
various
intervals
and
the
reaction
was
stopped
with
Sarkosyl.
Samples
were
then
subjected
to
PAGE
as
described
above.
RESULTS
The
MICs
and
MBCs
of
several
selected
P-lactam
anti-
biotics
for
L.
pneumophila
Philadelphia
strain
1
are
shown
in
Table
1.
Under
the
conditions
of
the
assay,
the
organisms
grew
very
slowly
in
both
antibiotic
and
control
tubes.
Therefore,
the
MIC
tubes
were
read
after
48
hand
again
after
72
h
of
growth.
For
all
antibiotics
tested
the
MBCs
differed
from
the
72-h
MICs
by
at
most
one
dilution
tube.
At
antibiotic
concentrations
of
1Ox
MIC,
there
was
promhpt
lysis
of
L.
pneumophila
cells
as
indicated
by
the
rapid
loss
of
culture
turbidity
(Table
2).
In
general,
benzylpenicillin
and
mecillinam
caused
the
most
rapid
loss
of
turbidity,
followed
by
thienamycin
and
then
aztreonam.
A
decrease
of
approx-
imately
50%
in
initial
turbidity
occurred
after
2
h
of
exposure
to
each
antibiotic
except
aztreonam.
The
latter
did
not
achieve
the
same
degree
of
lysis
until
after
4
h
of
exposure.
On
the
same
time
scale
the
lytic
response
of
L.
pneumophila
after
exposure
to
penicillin
was
delayed
in
onset
when
compared
with
that
of
E.
coli
cultures
(14).
Alteration
of
pH
of
the
medium
from
6.4
to
7.4
did
not
affect
the
ability
of
penicillin
G
to
promote
cell
lysis.
Below
pH
6.4
growth
of
the
organism
was
inadequate
to
allow
experimentation.
Loss
of
viability
of
L.
pneumophila
after
exposure
to
3-lactam
antibiotics
paralleled
the
decline
in
turbidity
(Fig.
1).
All
antibiotics
investigated
were
bactericidal.
Aztreonam
caused
a
slower
decline
in
CFUs
than
the
other
,B-lactams
tested,
and
a
99.9%
loss
of
viability
was
achieved
at
24
h
with
an
antibiotic
concentration
of
8x
MIC.
In
contrast,
penicillin
G
at
lx
MIC
achieved
a
comparable
decrease
in
viability
at
24
h.
At
sub-MICs
(e.g.,
0.5x
MIC)
we
often
observed
an
initial
decline
in
viability
at
24
h,
with
regrowth
of
the
culture
occurring
upon
longer
ihcubation.
When
tested,
these
organisms
had
the
same
MIC
as
the
original
inoculum,
an
indication
that
resistance
did
not
develop
after
penicillin
exposure
at
sub-MICs.
To
further
investigate
the
physiology
of
the
lytic
response
of
L.
pneumophila
to
P-lactam
antibiotics,
a
series
of
label-
ing
experiments
was
done.
Cells
were
labeled
with
[3H]uracil
and
then
exposed
to
lytic
doses
of
benzylpenicillin
(25
,tg/ml)
(Fig.
2).
Release
of
radioactively
labeled
nucleic
acids
into
the
medium
was
assayed
by
determining
trichlo-
roacetic
acid-precipitable
counts
in
the
supematant
after
centrifugation.
As
can
be
seen,
significant
amounts
of
nucl-
eic
acids
were
liberated
into
the
surrounding
medium
paral-
lel
with
the
decline
in
culture
turbidity.
Over
50%
of
the
macromolecular
uracil
label
was
released
from
the
penicil-
lin-treated
cells
in
3
h.
Despite
the
significant
loss
of
turbidity
and
viability
and
release
of
intracellular
macromolecules
during
penicillin-in-
duced
cell
lysis,
there
seemed
to
be
very
little
accompanying
cell
wall
degradation.
Figure
2
shows
the
amount
of
[3H]glucosamine
remaining
in
intact
cell
wall
polymers
(SDS-
insoluble
material)
during
a
4-h
exposure
to
benzylpenicillin.
There
was
little,
if
any,
decline
in
the
SDS-insoluble
fraction
of
radioactivity
after
penicillin-induced
lysis
of
L.
pneumo-
phila.
Subsequent
exposure
of
the
penicillin-treated
cells
to
Ml
muramidase
for
4
h
caused
an
80%
loss
of
SDS-insoluble
counts,
further
supporting
the
lack
of
cell
wall
degradation
by
penicillin
alone
(data
not
shown).
The
lack
of
peptidoglycan
degradation
during
penicillin-
induced
lysis
was
also
consistent
with
the
retention
of
0.5
1
2
4
8
0.5
1
2
4
8
0.5
1
2
4
8
Antibiotic
concentration
I
x
MIC
I
FIG.
1.
Bactericidal
activity
of
various
3-lactam
antibiotics.
Cultures
were
grown
at
37-C
with
5%
CO2
without
shaking.
In
the
exponential
growth
phase,
cultures
were
diluted
to
an
inoculum
concentration
of
3
x
10-
CFU/ml,
and
antibiotics
were
added
at
the
concentrations
(multiples
of
the
MIC)
indicated.
Viability
was
determined
as
described
in
the
text.
Symbols:
0,
benzylpenicillin;
0,
aztreonam;
L,
cefoxitin;
A,
thienamycin.
VOL.
27,
1985
697

698
WEISHOLTZ
AND
TOMASZ
bacillary
morphology
after
antibiotic
exposure.
Electron
micrographs
(not
shown)
of
L.
pneumophila
after
penicillin-
induced
lysis
demonstrated
only
a
general
decrease
in
cell
contents
(decrease
in
electron-dense
material)
that
corre-
sponded
to
the
decrease
in
turbidity
and
release
of
RNA
and
cell
contents
into
the
medium.
The
lack
of
correlation
between
cell
turbidity
and
intactness
of
peptidoglycan
was
also
demonstrated
in
experiments
in
which
[3H]glucosamine-
labeled
L.
pneumophila
was
treated
with
1%
SDS
or
with
Ml
muramidase.
In
the
former
case,
80%
of
turbidity
was
lost
without
any
solubilization
of
peptidoglycan
label.
In
contrast,
a
virtually
complete
degradation
of
peptidoglycan
by
Ml
muramidase
was
accompanied
by
only
about
a
20%
drop
in
turbidity.
Isolated
murein
sacculi
labeled
with
[3H]glucosamine
from
L.
pneumophila
appeared
to
be
susceptible
to
degradation
by
exogenous
muramidases
despite
the
inability
of
benzyl-
penicillin
to
induce
cell
wall
degradation
in
vivo.
Table
3
presents
the
results
of
overnight
peptidoglycan
digestion
by
various
murein
hydrolases.
As
can
be
seen,
the
N,O-diacetyl
muramidases,
such
as
the
Chalaropsis
muramidase
and
B
0
c
._
-W
-J
50j
r
1
2
3
Time
I
h
I
4
FIG.
2.
Penicillin-induced
lysis
of
L.
pneumophila
and
its
effect
on
release
of
cell
wall
and
intracellular
material.
(A)
Loss
of
turbidity
after
exposure
to
lytic
doses
of
benzylpenicillin
(25
,ug/ml)
was
determined
by
nephelometry.
Symbols:
0,
benzylpenicillin;
0,
control.
(B)
Release
of
[3H]uracil-labeled
nucleic
acids
from
the
intracellular
pool
and
[3H]glucosamine
label
from
the
SDS-insoluble
cell
wall
fraction
After
exposure
to
lytic
doses
of
benzylpenicillin
(25
,ug/ml)
was
determined
as
indicated
in
the
text.
Data
are
expressed
as
percentages
of
radioactive
label
remaining
in
the
intracellular
pool
or
SDS-insoluble
cell
wall
fraction,
respectively.
Each
point
represents
a
mean
of
two
experiments.
Symbols:
A,
[3H]uracil,
penicillin
exposure;
A,
[3H]uracil,
control;
0,
[3H]glucosamine,
penicillin
exposure;
0,
[3H]glucosamine,
control.
TABLE
3.
Hydrolysis
of
murein
sacculi
by
various
exogenous
muramidasesa
Murein
Muramidase
(,g/ml)
hydrolysis
(%)
Pure
Ml
muramidase
(10)
...................
95
Chalaropsis
enzyme
(10)
........
...........
98
N-Acetyl
muramidase
(100)
............6.......
9
Lysozyme
(100)
...................
64
Trypsin
(100)
0...................
a
[3H]glucosamine-labeled
cell
walls
(5
x
103
cpm
and
50
,g/100
,ul)
were
incubated
with
a
variety
of
enzymes
in
reaction
mixtures
of
1-ml
volumes
made
up
of
100
IL1
of
cell
wall
substrate
and
900
,u1
of
the
following
buffers:
100
mM
potassium
phosphate
at
pH
7.0
in
the
cases
of
Ml
and
N-acetyl
muramidase,
at
pH
5.0
for
the
Chalaropsis
enzyme,
and
at
pH
6.5
for
egg
white
lysozyme.
Tris
(pH
8.0;
100
mM)
was
used
with
trypsin.
The
mixtures
were
incubated
for
12
h
at
room
temperature,
and
the
fraction
of
SDS-
insoluble
counts
was
determined
as
described
in
the
text.
The
degree
of
murein
hydrolysis
was
expressed
in
relation
to
an
undigested
control.
purified
Ml
muramidase,
caused
a
95
to
98%
loss
of
SDS-
insoluble
counts,
indicating
near
total
degradation
of
the
murein
sacculi.
Trypsin
treatment
did
not
release
any
radio-
active
label
from
the
SDS-insoluble
murein,
and
egg
white
lysozyme
produced
only
a
64%
degriadation.
TLC
of
Chalaropsis-treated
murein
sacculi
confirmed
the
virtually
complete
degradation
of
the
L.
pneumophila
cell
wall
by
exogenous
muramidase
(Fig.
3).
A
similarly
digested
peptidoglycan
of
Neisseria
gonorrhoeae
was
also
chromato-
graphed
for
comparison
(courtesy
of
T.
Dougherty,
Rocke-
feller
University).
More
than
96%
of
the
original
radioactive
glucosamine
label
migrated
either
as
the
disaccharide
pep-
tide
monomer
or
as
the
bis-disaccharide
peptide,
and
no
O-acetylated
components
were
detectable
by
this
method.
To
further
elucidate
the
interactioh
of
1-lactam
antibiotics
with
L.
pneumophila,
we
investigated
the
PBPs
on
the
bacterial
inner
membranes.
After
in
vivo
labeling
of
log-
phase
cultures
with
tritiated
penicillin,
we
assayed
PBPs
in
an
SDS-PAGE
system.
Five
binding
proteins
were
consist-
ently
detected
(Fig.
4).
Their
respective
molecular
weights
FIG.
3.
TLC
of
the
[3H]glucosamine-labeled
murein
fragments
generated
by
digestion
of
L.
pneumophila
(Lp)
cell
walls
with
the
Chalaropsis
muramidase.
A
total
of
8,000
cpm
of
enzyme
hydroly-
sate
was
spotted
on
the
chromatogram.
Also
shown
are
the
muram-
idase
degradation
products
of
gonococcal
peptidoglycan
(Gc)
(5).
The
digested
fragments
include
mono-O-acetylated
monomers
(AcM),
monomers
(M),
di-O-acetylated
dimers
(diAcD)),
mono-0-
acetylated
dimers
(AcD),
and
dimers
(D).
ANTIMICROB.
AGENTS
CHEMOTHER.
I
0
2
u
0
_
_
Ad
_
__,&

LYSIS
WITHOUT
WALL
DEGRADATION
IN
L.
PNEUMOPHILA
PBP
1
2
3
4
5
FIG.
4.
PBPs
of
L.
pneumophila.
L.
pneumophila
cells
were
treated
with
6.25
,ug
of
[3H]penicillin
per
ml,
and
the
PBPs
were
separated
by
SDS-PAGE
followed
by
fluorography,
as
described
in
the
text.
The
fluorograms
were
scanned
with
a
Quick
Scan
Junior
densitometer
(Helena
Laboratories,
Beaumont,
Tex.).
were
95,000, 71,000,
61,000,
40,000,
and
19,000.
PBP
2
was
occasionally
resolved
into
two
bands
migrating
at
molecular
weights
of
-70,000
and
-64,000.
Pretreatment
of
the
cells
with
nonradioactive
penicillin,
Sarkosyl,
or
boiling
pre-
vented
binding
of
[3H]penicillin
to
these
binding
proteins.
Incubation
of
prelabeled
membrane
preparations
with
0.2
M
neutral
hydroxyamine
for
20
min
caused
the
loss
of
most
of
the
bound
penicillin
from
the
PBPs.
PBP
4
bound
the
greatest
amount
of
label
(57%),
followed
by
PBPs
1
and
2
(15
to
18%).
However,
PBPs
1
and
2
appeared
to
have
the
highest
affinity
for
[3H]penicillin,
with
labeling
at
a
concen-
tration
of
0.1
to
0.5
x
MIC
for
the
organism.
PBP
3
appeared
with
1
to
2x
MIC,
and
PBPs
4
and
5
did
not
appear
until
a
concentration
of
5
to
1Ox
MIC
was
used.
In
vitro
labeling
with
[3H]penicillin
of
membrane
prepara-
tions
from
L.
pneumophila
demonstrated
a
similar
PBP
pattern.
No
loss
of
label
from
any
of
the
penicilloyl-PBPs
could
be
demonstrated
during
60
min
after
postincubation
with
nonradioactive
penicillin.
DISCUSSION
The
L.
pneumophila
strain
examined
exhibited
several
of
the
typical
responses
shown
by
gram-negative
bacilli
treated
with
,-lactam
antibiotics.
The
organism
contained
five
PBPs
accessible
to
penicillin
both
in
membranes
and
in
the
grow-
ing
cell.
Treatment
with
aztreonam
or
mecillinam
resulted
in
a
tendency
to
form
elongated
or
spheroid
cells,
respectively
(undocumented
finding).
This
is
reminiscent
of
the
behavior
of
E.
coli
and
other
gram-negative
bacilli
and
suggests
the
presence
in
L.
pneumophila
of
PBPs
with
selective
affinities
for certain
P-lactam
antibiotics
(16).
Both
morphological
changes
required
prolonged
incubations
with
the
drugs,
presumably
because
of
the
long
doubling
time
(3
to
4
h)
of
these
bacteria
(7,
19).
Cell
walls
(sacculi
insoluble
in
hot
SDS)
of
this
L.
pneu-
mophila
strain
labeled
with
radioactive
glucosamine
could
be
fully
degraded
by
treatment
with
either
the
Streptomyces
globus
Ml
muramidase
or
the
Chalaropsis
muramidase
(5).
In
analysis
by
TLC
(10),
over
96%
of
the
degraded
material
was
found
to
have
migrated
as
a
mixture
of
the
disaccharide
peptide
monomer
(about
60%)
and
the
bis-disaccharide
peptide
dimer
(30%).
A
small
percentage
of
the
radioactivity
migrated
in
the
region
of
higher
oligomers.
There
was
no
evidence
for
O-acetylated
muramyl
derivatives.
With
the
data
from
two
representative
experiments,
the
percentage
of
cross-linking
between
stem-peptides
of
the
peptidoglycan
of
this
L.
pneumophila
strain
may
be
calculated
as
19
to
25%
(i.e.,
[(0.5
x
the
radioactivity
in
dimers)
+
(0.7
x
the
radioactivity
in
trimers]/total
radioactivity
recovered).
This
value
is
within
the
range
observed
in
other
gram-negative
peptidoglycans.
In
another
strain
of
L.
pneumophila
grown
in
a
different
medium
and
analyzed
by
a
different
technique,
Amano
and
Williams
obtained
results
that
were
interpreted
as
suggesting
a
much
higher
degree
of
peptidoglycan
cross-
linking
(1).
Philadelphia
strain
1
had
reasonably
low
MICs
and
MBCs
for
several
P-lactam
antibiotics,
and
the
cells
were
sensitive
to
both
the
cidal
and
lytic
effects
of
those
agents.
These
findings
together
support
the
notion
that
the
relative
resist-
ance
of
L.
pneumophila
infections
to
P-lactam
antibiotics
(2,
6,
9)
may
be
primarily
related
to
the
peculiar
mode
of
in
vivo
growth
of
this
bacterium
in
phagosomes
(12).
We
did
observe
one
unusual
feature
in
the
response
of
L.
pneumophila
to
P-lactam
antibiotics.
Treatment
of
L.
pneu-
mophila
cultures
with
penicillin
caused
a
rapid
loss
of
viability
and
lysis
of
bacteria
(as
determined
by
turbidity
drop
and
release
of
labeled
nucleic
acids
into
the
medium),
but
lysis
was
not
accompanied
by
cell
wall
degradation.
There
was
no
significant
decrease
of
the
SDS-insoluble
[3H]glucosamine-labeled
cell
wall
material
during
penicillin
exposure,
indicating
that
the
peptidoglycan
network
re-
mained
as
a
large
polymer
(Fig.
2).
In
contrast,
similar
antibiotic
treatments
of
pneumococci
or
E.
coli
resulted
in
the
loss
of
60
to
80%
of
the
original
SDS-insoluble
radioac-
tively
labeled
cell
wall
material
(results
not
shown).
The
morphological
appearance
of
L.
pneumophila
after
treat-
ment
with
penicillin
also
suggested
the
relative
integrity
of
the
murein
sacculus
despite
a
general
"fading"
of
cyto-
plasm.
On
the
other
hand,
the
cell
envelope
of
E.
coli
became
fragmented
and
disrupted
after
exposure
to
penicil-
lin
(14).
Possibly,
the
1-lactam
antibiotic-induced
lysis
of
L.
pneumophila
involves
a
murein
hydrolase
that
nicks
but
does
not
degrade
the
peptidoglycan.
In
this
respect,
it
is
interesting
that
treatment
of
the
L.
pneumophila
cell
wall
with
egg
white
lysozyme
has
been
reported
to
yield
virtually
quantitative
cleavage
of
the
glycan
to
disaccharides
without
solubilization
of
the
wall
(1).
It
has
been
shown
repeatedly
that
the
physiological
effects
of
,-lactam
antibiotics
on
microorganisms
vary
greatly
with
the
target
organism.
In
tolerant
strains
of
Streptococcus
sanguis
which
lack
autolytic
activity,
treatment
with
peni-
cillin
causes
inhibition
of
growth
but
no
lysis,
no
cell
wall
degradation,
and
only
very
slow
loss
of
viability.
In
Strepto-
coccus
pyogenes,
penicillin
causes
a
halt
in
growth
and
rapid
loss
of
viability
but
no
lysis
or
cell
wall
degradation
(11).
In
E.
coli
(14)
or
pneumococci
(18),
treatment
with
inhibitors
of
cell
wall
synthesis
causes
rapid
killing,
cell
lysis,
and
cell
wall
degradation
as
well.
The
novel
type
of
cell
lysis
without
cell
wall
degradation,
as
described
in
this
report
for
L.
pneumophila,
appears
to
be
yet
another
example
of
the
variation
in
the
physiological
responses
of
bacteria
to
,B-
lactam
antibiotics.
ACKNOWLEDGMENTS
We
acknowledge
the
assistance
of
Thomas
Dougherty.
These
investigations
were
supported
by
a
Fellowship
in
Infectious
Diseases
from
Cornell
University
Medical
School
(to
S.W.).
LITERATURE
CITED
1.
Amano,
K.,
and
J.
C.
Williams.
1983.
Peptidoglycan
of
Legion-
ella
pneumophila:
apparent
resistance
to
lysozyme
hydrolysis
correlates
with
a
high
degree
of
peptide
cross-linking.
J.
Bac-
teriol.
153:520-526.
2.
Bgcheson,
M.
A.,
H.
M.
Friedman,
and
C.
E.
Benson.
1981.
Antimicrobial
susceptibility
of
intracellular
Legionella
pneu-
mophila.
Antimicrob.
Agents
Chemother.
20:691-692.
699
VOL.
27,
1985

700
WEISHOLTZ
AND
TOMASZ
3.
Blundell,
J.
K.,
G.
R.
Smith,
and
H.
R.
Perkins.
1980.
The
peptidoglycan
of
Neisseria
gonorrhoeae:
O-acetyl
groups
and
lysozyme
sensitivity.
FEMS
Microbiol.
Lett.
9:259-261.
4.
Buchan,n,
C.
E.,
and
J.
L.
Strominger.
1976.
Altered
penicil-
lin-binding
components
in
penicillin-resistant
mutants
of
Bacil-
lus
subtilis.
Proc.
Natl.
Acad.
Sci.
U.S.A.
73:1816-1820.
5.
Dougherty,
T.
J.
1983.
Peptidoglycan
biosynthesis
in
Neisseria
gonorrhoeae
strains
sensitive
and
intrinsically
resistant
to
13-
lactam
antibiotics.
J.
Bacteriol.
153:429-435.
6.
Edelstein,
P.
H.,
a,nd
R.
D.
Meyer.
1980.
Susceptibility
of
Legionella
pneumophila
to
twenty
antimicrobial
agents.
Antimi-
crob.
Agents
Chemother.
18:403408.
7.
Feeley,
J.
C.,
G.
W.
Gorman,
R.
E.
Weaver,
D.
C.
Mackel,
and
H.
W.
Smith.
1978.
Primary
isolation
media
for
Legionnaires
disease
bacterium.
J.
Clin.
Microbiol.
8:320-325.
8.
Flesher,
A.
R.,
S.
Ito,
B.
J.
Monsheim, and
D.
L.
Kasper.
1979.
The
cell
envelope
of
the
Legionnaire's
bacterium-morphologic
and
biochemical
characteristics.
Ann.
Intern.
Med.
90:628-631.
9.
Fraser,
D.
W.,
C.
Bapp,
I.
I.
Wachsmuth,
S.
C.
Feeley,
and
T.
F.
Tsai.
1978.
Antibiotic
treatment
of
guinea
pigs
infected
with
agents
of
Legionnaire's
disease.
Lancet
i:175-177.
10.
Gmeiner,
J.,
and
H.-P.
Kroll.
1981.
Murein
biosynthesis
and
O-acetylation
of
N-acetylmuramic
acid
during
the
cell-division
cycle
of
Proteus
mirabilis.
Eur.
J.
Biochem.
117:171-177.
11.
Horne,
D.,
and
A.
Tomasz.
1977.
Tolerant
response
of
Stepto-
coccus
sanguis
to
beta-lactams
and
other
cell
wall
inhibitors.
Antimicrob.
Agents
Chemother.
11:888-896.
12.
Horwitz,
M.
A.,
and
C.
A.
Silverstein.
1980.
Legionnaire's
disease
bacterium
(Legionella
pneumophila)
multiplies
intracel-
lularly
in
human
monocytes.
J.
Clin.
Invest.
66:441-450.
13.
Keel,
J.
A.,
W.
R.
Finpely,
and
J.
C.
Feeley.
1979.
Fine
structure
of
the
Legionnaire's
bacterium-in
vitro
and
in
vivo
studies
of
four
isolates.
Ann.
Intern.
Med.
90:652-656.
14.
Kitano,
K.,
and
A.
Tomasz.
1979.
Triggering
of
autolytic
cell
wall
degradation
in
Escherichia
coli
by
beta-lactam
antibiotics.
Antimicrob.
Agents
Chemother.
16:838-848.
15.
Lacks,
S.,
and
R.
D.
Hotchkiss.
1960.
A
study
of
the
genetic
material
determining
an
enzyme
activity
in
pneumococcus.
Biochim.
Biophys.
Acta
106:564-576.
16.
Spratt,
B.
G.
1977.
Properties
of
the
penicillin
binding
proteins
of
Escherichia
coli.
Eur.
J.
Biochem.
72:341-352.
17.
Tomasz,
A.,
J.
D.
Jamieson,
and
E.
Ottolenghi.
1964.
The
fine
structure
of
Diplococcus
pneumoniae.
J.
Cell
Biol.
22:453-480.
18.
Tom4sz,
A.,
and
S.
Waks.
1975.
Mechanism
of
action
of
penicillin:
triggering
of
the
pneumococcal
autolytic
enzyme
by
inhibitors
of
cell
wall
synthesis.
Proc.
Natl.
Acad.
Sci.
U.S.A.
72:41624166.
19.
Warren,
W.
J.,
and
R.
J.
Miller.
1979.
Growth
of
Legionnaires
disease
bacterium
(Legionella
pneumophila)
in
chemically
de-
fined
medium.
J.
Clin.
Microbiol.
10:50-55.
ANTIMICROB.
AGENTS
CHEMOTHER.