ArticlePDF Available

Role of β-lactam hydrolysis in the mechanism of resistance of a β-lactamase-constitutive Enterobacter cloacae strain to expanded-spectrum β-lactams

Authors:

Abstract

Enterobacter cloacae strains producing chromosomally mediated beta-lactamase constitutively show high degrees of resistance to most of the third-generation beta-lactams. It has been proposed that this resistance is due to the nonhydrolytic binding or trapping of beta-lactams by the enzyme. We found that the outer membrane of E. cloacae strain 55M indeed had permeability to cefazolin about 14-fold lower than that of Escherichia coli, and that the number of beta-lactamase molecules produced by this constitutive mutant was exceptionally large (2 X 10(5) per cell). These conditions are expected to produce a low degree of resistance, but could not explain the high resistance level of the mutant. We showed that the beta-lactamase of this strain hydrolyzed third-generation beta-lactams at measurable rates. Although the V max for these compounds was less than 0.01% of that for cefazolin, the enzyme could hydrolyze them at rates comparable to the rate for cefazolin when the substrate concentration was near 0.1 microM, a concentration thought to be physiologically relevant for the inhibition of cell growth, because of the exceptionally high affinity of the enzyme to many third-generation compounds. Calculations based on kinetic parameters of the enzyme, outer membrane permeability, and affinity toward penicillin-binding proteins succeeded in predicting the MICs for several third-generation beta-lactams. The data suggest that hydrolysis may be more important than nonhydrolytic binding for the expression of the resistant phenotype, and that studies on the susceptibility of beta-lactams to beta-lactamases should be carried out at physiologically relevant, very low concentrations of the drug, rather than the customary very high concentrations, such as 100 microM.
Vol.
27,
No.
3
ANTIMICROBIAL
AGENTS
AND
CHEMOTHERAPY,
Mar.
1985,
p.
393-398
0066-4804/85/030393-06$02.00/0
Copyright
C)
1985,
American
Society
for
Microbiology
Role
of
,B-Lactam
Hydrolysis
in
the
Mechanism
of
Resistance
of
a
3-Lactamase-Constitutive
Enterobacter
cloacae
Strain
to
Expanded-
Spectrum
r-Lactams
HANH
VU
AND
HIROSHI
NIKAIDO*
Department
of
Microbiology
and
Immunology,
University
of
California,
Berkeley,
California
94720
Received
10
September
1984/Accepted
7
December
1984
Enterobacter
cloacae
strains
producing
chromosomally
mediated
j8-lactamase
constitutively
show
high
degrees
of
resistance
to
most
of
the
third-generation
I8-lactams.
It
has
been
proposed
that
this
resistance
is
due
to
the
nonhydrolytic
binding
or
trapping
of
f-lactams
by
the
enzyme.
We
found
that
the
outer
membrane
of
E.
cloacae
strain
55M
indeed
had
permeability
to
cefazolin
about
14-fold
lower
than
that
of
Escherichia
coli,
and
that
the
number
of
I-lactamase
molecules
produced
by
this
constitutive
mutant
was
exceptionally
large
(2
x
105
per
cell).
These
conditions
are
expected
to
produce
a
low
degree
of
resistance,
but
could
not
explain
the
high
resistance
level
of
the
mutant.
We
showed
that
the
P-lactamase
of
this
strain
hydrolyzed
third-generation
j8-lactams
at
measurable
rates.
Although
the
Vmax
for
these
compounds
was
less
than
0.01%
of
that
for
cefazolin,
the
enzyme
could
hydrolyze
them
at
rates
comparable
to
the
rate
for
cefazolin
when
the
substrate
concentration
was
near
0.1
,uM,
a
concentration
thought
to
be
physiologically
relevant
for
the
inhibition
of
cell
growth,
because
of
the
exceptionally
high
affinity
of
the
enzyme
to
many
third-generation
compounds.
Calculations
based
on
kinetic
parameters
of
the
enzyme,
outer
membrane
permeability,
and
affinity
toward
penicillin-binding
proteins
succeeded
in
predicting
the
MICs
for
several
third-generation
P-lactams.
The
data
suggest
that
hydrolysis
may
be
more
important
than
nonhydrolytic
binding
for
the
expression
of
the
resistant
phenotype,
and
that
studies
on
the
susceptibility
of
j-lactams
to
P-lactamases
should
be
carried
out
at
physiologically relevant,
very
low
concentrations
of
the
drug,
rather
than
the
customary
very
high
concentra-
tions,
such
as
100
,IM.
In
recent
years
we
have
witnessed
the
isolation,
both
from
clinical
sources
and
in
the
laboratory,
of
strains
of
gram-neg-
ative
bacteria,
especially
Enterobacter
cloacae,
that
are
resistant
to
a
number
of
the
third-generation
P-lactams
(15).
These
strains
usually
show
the
following
characteristics
(15):
(i)
they
produce,
in
a
constitutive
manner,
the
chromosom-
ally
determined
,-lactamase
that
is
produced
inducibly
in
the
wild-type
cells;
and
(ii)
the
spectrum
of
the
third-gener-
ation
P-lactams
to
which
they
become
resistant
includes
many
agents
that
their
,B-lactamase
does
not
seem
to
hydro-
lyze
with
measurable
rates.
To
explain
this
puzzling
observation,
it
has
been
proposed
that
the
periplasmic
,B-lactamase
produces
resistance
by
tightly
combining
with
the
third-generation
,-lactams
with-
out
hydrolyzing
them
(15,
18-20).
This
concept
of
trapping
or
nonhydrolytic
barrier
has
been
challenged
on
the
ground
that,
for
this
mechanism
to
work,
the
total
number
of
3-lactamase
molecules
per
milliliter
of
culture
must
be
larger
than
the
number
of
P-lactam
molecules
in
the
same
volume
(17).
However,
this
argument
fails
to
take
into
account
the
presence
of
an
outer
membrane
barrier
and
the
fact
that,
during
any
given
time
period,
only
a
very
small
fraction
of
the
3-lactam
molecules
present
in
the
medium
passes
through
this
barrier
and
contacts
the
periplasmic
,B-lactamase.
In
this
study
we
have
examined,
in
a
quantitative
way,
the
extent
of
this
barrier,
the
number
of
,-lactamase
molecules
available
for trapping,
and
the
possibility
of
slow,
but
significant,
rates
of
hydrolysis
of
the
third-generation
agents
by
the
,B-
lactamase
to
determine
whether
the
resistance
is
truly
caused
by
a
trapping
mechanism.
*
Corresponding
author.
MATERIALS
AND
METHODS
Bacterial
strains.
E.
cloacae
55W
and
its
P-lactamase-
constitutive
mutant
55M
were
obtained
from
Christine
C.
Sanders
(5).
They
were
grown
in
L
broth
(10
g
of
Bacto-
tryptone
[Difco
Laboratories],
10
g
of
Bacto
yeast
extract
[Difco],
and
5
g
of
NaCl
per
liter)
at
37°C
with
aeration
by
shaking
and
were
harvested
when
the
Klett
reading
(red
filter)
reached
100
(about
0.26
mg
[dry
weight]
per
ml).
Chemicals.
The
1-lactams
used
were
donated
by
the
following
companies:
cefoxitin
and
imipenem
(Merck
Sharpe
&
Dohme),
azthreonam
(E.
R.
Squibb
&
Sons),
ceftriaxone
(Hoffmann-La
Roche
Inc.),
cefotaxime
(Hoechst-Roussel),
ceftazidime
(Glaxo),
ceftizoxime
(Fujisawa),
and
cefoper-
azone
(Pfizer
Inc.).
Cefazolin
was
obtained
from
Sigma
Chemical
Co.
Determination
of
MIC.
One
drop
of
a
suspension
contain-
ing
105
cells
per
ml
was
deposited
on
the
surface
of
L-agar
plates
containing
serial
twofold
dilutions
of
,-lactams,
and
the
growth
was
examined
after
overnight
incubation
at
37°C.
Growth
of
fewer
than
30
colonies
was
scored
as
negative.
Purification
of
j3-lactamase
by
high-pressure
liquid
chroma-
tography-gel
filtration.
Strain
55M
was
grown
in
minimal
medium
63
(2)
and
was
subjected
to
osmotic
shock
as
described
previously
(14).
Concentrated
osmotic
shock
su-
pernatant
was
injected
into
a
TSK
3000SW
column
(6.5
by
600
mm,
Beckman
Instruments,
Inc.)
connected
to
a
Perkin-
Elmer
LC-75
variable-wavelength
detector
and
a
Beckman
model
110A
pump,
and
gel
filtration
was
performed
with
0.1
M
Na2SO4-0.01
M
sodium
phosphate
buffer
(pH
7.0)
pumped
at
a
flow
rate
of
1.0
ml
min-'.
The
effluent
was
monitored
at
280
nm.
Fractions
were
collected
every
30
s
and
were
analyzed
for
protein
composition
by
sodium
dodecyl
sulfate
393
ANTIMICROB.
AGENTS
CHEMOTHER.
TABLE
1.
Catalytic
properties
of
the
E.
cloacae
55
M
1-lactamase
and
their
effect
on
resistance
V
at
0.1
I±M
C,,
pro-
,13-Lactam
V,,,,
(nmol
mg-'
(molecules
Permeability
ducing
MiC
P-Lactam
~
per
s)
K.,
(4RM)
K,
(LmM)
cefll'
coefficienta
Cp
of
(~Lgml)
per
s)
(nm
s'1)
0.1
;LMG
pers)
~~~~~~~(Rg/mI)
Cefazolin
2300
(100)b
2080
34,600
35
114
>250
Cefoperazone
2.7
(0.12)
6.5c
2.1d
12,800
7
295
125
Ceftazidime
0.05
(0.002)
3.8'
2.6d
430
5
12
125
Ceftriaxone
NDe
ND
0.03f
250
Cefoxitin
0.06
(0.002)
<0.059
0.023h
14,000
20
76
125
Ceftizoxime
0.008
(0.0004)
0.3
0.12
620
15
4
125
Cefotaxime
0.023
(0.001)
<0.05k
0.003'
5,620
9
73
250
Azthreonam
ND ND
0.012'
25
a
See
the
text
for
methods
of
calculation.
"Values
within
parentheses
indicate
relative
rates,
with
the
rate
for
cefazolin
set
at
100.
C
These
values
are
similar
to
16
and
6.6
FLM
for
cefoperazone
and
ceftazidime,
respectively,
reported
for
E.
cloacae
P99
enzyme
(1).
d
These
values
are
similar
to
9
and
3
FM
for
cefoperazone
and
ceftazidime,
respectively,
reported
for
E.
cloacae
P99
enzyme
(1).
'
ND,
Not
determined.
f
A
fairly
close
value
of
0.05
to
0.09
FtM
has
been
reported
(19).
'
These
Km
values
were
approximated
from
the
time
course
of
complete
hydrolysis
of
1
p.M
substrates.
When
higher
concentrations
of
substrates
were
used,
higher
apparent
Km
values
in
the
range
of
2.5
to
5
F.M
were
suggested
from
Lineweaver-Burk
plots,
as
described
in
text.
h
Reported
values
are
0.12
to
1.25
,uM
(19)
for
various
strains
and
0.5
F.M
for
GN7471
(10).
'Reported
values
are
0.04
to
0.07
F.M
(19),
0.05
,uM
(10),
and
0.035
FM
(1).
'
K1
of
0.012
pM
has
been
reported
(1).
(SDS)-polyacrylamide
gel
electrophoresis
(see
below)
and
for
,B-lactamase
activity
with
cefazolin.
Assay
of
permeability.
The
permeability
assay
was
carried
out
by
the
method
of
Zimmermann
and
Rosselet
(22)
by
measuring
the
rate
of
hydrolysis
of
a
P-lactam,
usually
cefazolin,
by
intact
cells
as
well
as
by
sonic
extracts
of
these
cells,
and
calculating
the
permeability
of
the
outer
mem-
brane
on
the
basis
of
the
Km
(2.08
mM
for
cefazolin
in
strain
55M)
and
Vmx
of
the
enzyme
(13).
The
hydrolysis
rates
by
intact
cells
were
corrected
for
the
contribution
by
extracel-
lular
enzyme,
but
this
correction
amounted
to
less
than
0.5%
of
the
intact
cell
rates.
Assay
of
1-lacamase.
The
3-lactamase
assay
was
carried
out
spectrophotometrically
by
following
the
decrease
of
the
absorption
peak
around
260
nm
in
10
mM
potassium
phos-
phate
buffer
(pH
7).
The
exact
wavelengths
and
the
molar
differential
absorbance
values
used
were
those
listed
in
reference
11
for
cefazolin,
cefotaxime,
ceftizoxime,
cefoper-
azone,
and
cefoxitin.
As
reported
by
Bush
et
al.
(1),
hydro-
lysis
of
azthreonam
produces
maximum
absorbance
changes
around
318
nm,
but
the
magnitude
of
the
change
was
too
small
to
be
useful
for
our
assay.
Seeberg
et
al.
(17)
recom-
mend
272
nm
for
ceftazidime,
but
we
found
that
the
optical
density
changes
were
minimal
at
this
wavelength
and
that
they
were
maximal
at
257
nm;
we
used
the
latter
wavelength
for
assay.
For
ceftriaxone,
Seeberg
et
al.
(17)
recommend
240
nm,
and
we
confirmed
that
the
maximal
absorbance
change
occurred
at
236
nm.
However,
the
absorbance
in-
creased,
rather
than
decreased,
upon
hydrolysis,
and
the
extent
of
change
was
so
small
that
it
was
not
useful
for
the
assay
of
the
enzymatic
hydrolysis
of
this
compound.
To
detect
the
very
slow
hydrolysis
of
some
of
the
agents
accurately,
we
used
spectrophotometers
(Beckman
DU-7
and
Hitachi
model
220)
that
did
not
show
much
noise
even
when
the
full
scale
was
expanded
to
cover
the
range
of
0.01
optical
density
unit.
The
assay
was
run
at
room
temperature,
and
the
results
were
recorded
usually
for
10
min
with
the
activity
determined
from
the
initial
slope
of
the
tracing.
For
determination
of
Ki
by
Dixon
plots,
rates
of
hydrolysis
of
100
,uM
cefazolin
were
determined
in
the
presence
of
various
concentrations
of
inhibitors.
In
these
experiments
0.5
nM
enzyme
was
preincubated
with
the
inhibitor
for
2
min
before
the
addition
of
cefazolin.
For
determination
of
the
approxi-
mate
Km
values
in
the
submicromolar
range,
we
used
continuous
tracing
of
the
process
of
complete
hydrolysis
of
1
,uM
substrate
by
the
enzyme
(usually
20
nM),
as
described
below.
Other
methods.
Protein
was
determined
by
the
Lowry
et
al.
method
(9),
and
SDS-polyacrylamide
gel
electrophoresis
was
carried
out
with
the
slab
version
of
the
Laemmli
method
(7).
Scanning
of
the
stained
gel
was
performed
with
a
Quick
Scan
gel
scanner
from
Helena
Laboratories.
RESULTS
MIC
values
for
various
-lactams.
We
could
confirm
that
the
,B-lactamase-constitutive
strain,
55M,
was
highly
resist-
ant
to
most
of
the
third-generation
3-lactams
(Table
1).
The
strain
remained
very
susceptible
to
imipenem
(MIC,
<0.5
,ug
ml-;
data
not
shown).
Number
of
,B-lactamase
molecules
per
cell.
When
the
cells
of
the
constitutive
mutant
55M
and
its
inducible
parent
strain
55W
were
solubilized
by
heating
at
100°C
for
2
min
in
the
Laemmli
sample
buffer
(7)
containing
SDS
and
mercapto-
ethanol,
and
the
solubilized
samples
were
analyzed
by
SDS-polyacrylamide
slab
gel
electrophoresis,
the
constitu-
tive
strain
was
seen
to
contain
a
prominent
band
with
an
apparent
molecular
weight
of
42,000
which
was
absent
in
strain
55W
(Fig.
1).
To
confirm
the
identification
of
this
band
as
the
-
lactamase,
cells
of
strain
55M
harvested
at
the
midexponen-
tial
phase
were
exposed
to
osmotic
shock
by
the
method
of
Nossal
and
Heppel
(14).
The
supernatant
after
the
shock
contained
more
than
95%
of
the
1-lactamase
activity
present
in
the
original
cells
and
contained
the
42,000-dalton
protein
as
the
most
prominent
component
observed
upon
SDS-po-
lyacrylamide
gel
analysis.
When
the
supernatant
was
con-
centrated
by
ultrafiltration
with
a
Millipore
CX
ultrafilter
and
was
fractionated
by
high-pressure
gel
filtration
(see
above),
the
peak
fraction
containing
3-lactamase
activity
was
found
to
contain
mostly
the
42,000-dalton
protein
(Fig.
1).
When
the
amount
of
the
42,000-dalton
band
was
quanti-
tated
by
scanning
the
Coomassie
blue-stained
gel
of
the
394
VU
AND
NIKAIDO
ENZYMATIC
HYDROLYSIS
OF
EXPANDED-SPECTRUM
1-LACTAMS
whole
cell
proteins
from
strain
55M,
it
corresponded
to
3.7%
of
the
total
cellular
protein.
Assuming
that
there
are
2
x
109
cells
per
mg
(dry
weight),
and
that
70%
of
the
dry
weight
is
protein,
this
means
that
there
are
1.9
x
105
P-lactamase
molecules
per
cell
of
the
constitutive
strain,
55M.
Permeability
of
E.
cloacae
outer
membrane
to
cefazolin.
When
exponential-phase,
washed
cells
of
strain
55M
grown
in
L
broth
were
incubated
in
the
presence
of
1
mM
cefazolin
and
the
rate
of
hydrolysis
was
determined
spectrophotometri-
cally
with
a
cell
of
1-mm
optical
path
(13),
the
rate
was
only
30
nmol
per
mg
(dry
weight)
per
min,
in
comparison
with
the
rate
obtained
with
the
sonic
extract
of
these
cells,
44,200
nmol
per
mg
(dry
weight)
per
min
in
the
presence
of
the
same
cefazolin
concentration.
From
the
Km
of
the
periplasmic
P-lactamase
for
cefazolin
(see
above)
and
the
area
of
the
cell
surface
per
milligram
(dry
weight)
(13),
we
calculated
that
the
permeability
coefficient
of
the
outer
membrane
of
this
strain
to
cefazolin
was
3.5
x
106
cm
s-1,
14
times
less
than
the
permeability
of
OmpF-containing
Escherichia
coli
cells
to
this
compound
(13).
Activity
of
E.
cloacae
P-lactamase
against
various
-
lactams.
The
kinetics
of
hydrolysis
of
several
,-lactams
was
determined
by
using
osmotic
shock
fluid
from
E.
cloacae
55M
as
the
enzyme
source.
All
of
the
third-generation
compounds
we
could
test
with
good
sensitivity
were
hydro-
lyzed
with
low,
but
measurable,
rates
(Table
1).
We
could
usually
observe
a
finite
Km
value
in
the
neighborhood
of
several
micromolar
by
using
substrate
concentration
range
of
2
to
50
,uM.
However,
we
do
not
believe
that
these
values
reflected
the
true
Km
values
for
the
third-generation
com-
pounds
except
for
cefoperazone
and
ceftazidime.
First,
the
71
1j
010
1);SAX,$:.0200000
FIG.
1.
SDS-polyacrylamide
slab
gel
electrophoresis
of
fraction-
ated
osmotic
shock
fluids.
Osmotic
shock
supernatant
from
strain
55M
was
applied
to
a
high-pressure
liquid
chromatograph
gel
filtration
column
as
described
in
the
text,
and
the
fractions
were
analyzed
by
slab
gel
electrophoresis.
Only
fraction
15,
which
contained
strong
P-lactamase
activity,
and
fraction
13,
devoid
of
the
activity,
are
shown
here.
Fraction
15
is
seen
to
contain
mostly
a
42,000-dalton
protein,
although
it
is
contaminated
by
the
trailing
edge
of
a
75,000-dalton
protein
peak,
some
of
which
is
seen
in
fraction
13.
The
two
lanes
on
the
right
show
the
protein
composition
of
whole
cells
of
55M
and
55W;
the
only
difference
seen
is
the
presence
and
absence
of
the
42,000-dalton
protein.
Ki
values
obtained
by
measuring
the
rate
of
hydrolysis
of
100
,uM
cefazolin
in
the
presence
of
these
compounds
were
orders
of
magnitude
lower
than
the
observed
Km
values,
with
the
exception
of
ceftazidime
and
cefoperazone.
Sec-
ond,
the
rate
of
hydrolysis
of
low
concentrations
of
these
compounds
other
than
cefoperazone
and
ceftazidime
(for
example,
a
1
,uM
solution)
was
constant
until
very
shortly
before
the
substrate
became
exhausted,
a
result
showing
that
the
true
Km
was
far
lower
than
1
puM.
Third,
with
cefotaxime,
it
was
clear
that
the
hydrolysis
had
two
compo-
nents,
one
with
very
low
Km
and
the
other
with
the
previ-
ously
measured
4.8
,uM
Ki,,
and
that
the
first
process
became
inhibited
at
higher
(i.e.,
>2
,uM)
substrate
concen-
trations,
because
the
hydrolysis
rate
was
about
twice
as
high
with
1
p.M
substrate
as
with
2.5
p.M
substrate.
Because
of
these
observations,
we
determined
the
true
Km
values
by
following
the
kinetics
of
hydrolysis
of
1
p.M
solutions
of
P-lactams
to
completion
and
by
reading
the
concentration
at
which
the
rate
became
one-half
of
the
initial
rate.
For
example,
if
the
optical
density
decreased
from
0.188
to
0.180
during
the
course
of
hydrolysis,
and
if
the
slope
became
one-half
of
the
initial
slope
at
the
optical
density
of
0.
181,
the
Km
was
1
x
(0.181
-
0.180)/(0.188
-
0.180)
=
0.125
p.M.
This
procedure
produced
Km
values
that
were
usually
only
slightly
higher
than
Ki
values
for
the
respective
P-lactams
(Table
1),
and
which
we
believe
represent
the
true
Km
values.
We
do
not
know
what
produces
the
high
Km
compo-
nents
observed.
It
may
be
that
substrate
inhibition
at
these
relatively
high
substrate
concentrations
produced
a
bend
in
the
Lineweaver-Burk
plot,
or
that
another
enzyme
exists
in
our
preparation
because
we
used
a
crude
osmotic
shock
fluid
for
the
reaction.
In
any
case
we
do
not
believe
that
the
high
Km
components
are
relevant
in
the
physiology
of
bacteria
living
in
the
presence
of
P-lactam,
as
described
below.
Vmax
values
obtained
for
the
third-generation
agents
were
low,
but
significant,
in
all
cases.
We
believe
that
past
studies
that
found
these
agents
to
be
non-hydrolyzable
failed
to
recognize
this
fact
because
of
the lack
of
sensitivity
of
the
methods
employed.
Indeed,
in
addition
to
cefoperazone,
which
is
generally
recognized
as
hydrolyzable,
slow
hydro-
lysis
of
cefoxitin
(10,
16)
as
well
as
cefotaxime
and
ceftazid-
ime
(1)
has
been
reported
previously.
The
very
low
levels
of
Vmax
for
the
third-generation
compounds
may
nevertheless
be
quite
significant
in
a
physiological
context,
as
described
fully
below.
But
we
can
already
calculate,
by
using
the
Vmax
and
Km
values
in
Table
1,
that,
although
there
will
be
a
more
than
1,000-fold
difference
in
the
hydrolysis
rates
of
cefazolin
and
cefoxitin
at
0.1
mM
substrate
concentration,
the
differ-
ence
will
be
reduced
to
about
20-fold
at
1
p.M
and
to
only
about
2-fold
at
0.1
p.M.
Indeed,
the
assay
of
cefoxitin
and
cefazolin
hydrolysis
rates
at
1
p.M
concentration
with
the
same
amount
of
enzyme
showed
a
difference
of
26-fold,
close
to
the
predicted
value
(Fig.
2).
DISCUSSION
It
has
been
proposed
by
several
laboratories
that
the
tight
binding
of
non-hydrolyzable
P-lactams
by
periplasmic
P-lactamase
molecules,
i.e.,
trapping
of
P-lactams,
could
create
resistance
to
third-generation
agents
seen
in
p-
lactamase-constitutive
mutants
of
E.
cloacae
(see
above).
Furthermore,
Seeberg
et
al.
(17)
ruled
out
the
possibility
that
other
changes
accompanying
the
production
of
the
peri-
plasmic
P-lactamase
are
the
true
cause
of
the
resistance.
However,
for
the
trapping
mechanism
to
work,
the
organism
has
to
possess
a
very
large
number
of
P-lactamase
molecules
395
VOL.
27,
1985
ANTIMICROB.
AGENTS
CHEMOTHER.
0
(0
0
CEFAZOLIN
0
2
3
4
5min
FIG.
2.
Hydrolysis
of
cefazolin
and
cefoxitin
at
1
,uM
substrate
concentration.
Osmotic
shock
fluid
corresponding
to
1.2
x
107
cells
was
added
to
1-ml
solutions
containing
1
,uM
substrate
and
10
mM
potassium
phosphate
buffer,
pH
7.0,
and
the
hydrolysis
of
the
substrate
was
followed
by
observing
optical
density
(OD)
at
260
nm
at
room
temperature
in
a
Beckman
DU-7
spectrophotometer.
The
curves
were
smoothed
out
to
eliminate
the
effect
of
noise.
and
at
the
same
time
allow
only
a
very
small
number
of
,-lactam
molecules
to
trickle
into
the
periplasm.
To
our
knowledge,
quantitative
estimation
of
neither
of
these
pa-
rameters
has
been
reported
previously.
In
this
study
we
determined
these
parameters
by
using
one
such
constitutive
mutant,
E.
cloacae
55M.
By
identifying
the
P-lactamase
band
on
SDS-polyacrylamide
slab
gel
and
quan-
titating
this
band
on
gels
of
whole
cell
proteins,
we
could
determine
that
the
strain
produced
a
very
large
amount
of
this
enzyme,
corresponding
to
3.7%
of
the
total
cellular
protein,
or
about
2
x
105
molecules
per
cell.
This
number
is
similar
to
the
number
of
another
very
abundant
periplasmic
protein
of
E.
coli,
maltose-binding
protein,
which
is
reported
to
be
present
in
4
x
105
copies
per
cell
(4).
If
the
cells
are
growing
with
a
generation
time
of
20
min,
a
cell
should
be
synthesizing,
on
an
average,
(2
x
105)/(20
x
60)
=
167
molecules
of
new
P-lactamase
every
second.
The
outer
membrane
permeability
to
cefazolin
was
low,
estimated
to
be
about
7%
of
the
permeability
coefficient
previously
determined
in
E.
coli
K-12
(5
x
10-5
cm/s
[13]).
Since
most
of
the
third-generation
,-lactams,
with
the
nota-
ble
exception
of
imipenem,
have
permeability
coefficients
lower
than
20%
of
that
of
cefazolin
(21),
and
since
the
major
porin
of
E.
cloacae
appears
to
have
a
pore
size
slightly
narrower
than
the
E.
coli
porins
(6,
12),
it
is
reasonable
to
assume
that
the
permeability
coefficients
of
the
E.
cloacae
for
them
would
be
lower
than
5
x
10-5
x
0.07
x
0.2
=
7
x
10-7
cm
s-1
(see
also
reference
13).
The
maximal
rate
of
influx
of
such
agents
across
the
outer
membrane
is
v
=
P
x
A
x
C0,
following
Fick's
first
law
of
diffusion,
where
P,
A,
and
C0
denote
the
permeability
coefficient,
area
of
the
membrane
per
unit
weight
of
cells,
and
the
external
concen-
tration
of
the
P-lactam,
respectively.
Using
the
value
of
132
cm2
mg-1
for
A
(13),
and
assuming
C0
=
0.4
x
10-8
mol
cm-3
or
about
2
,ug
ml-'
with
a
1-lactam
molecular
weight
of
500,
v
=
7
x
10-7
(cm
s-1)
x
132
(cm2
mg-1)
x
0.4
x
10-8
(mol
cm-3)
x
6.023
x
1023
(molecules
mol-')/[2
x
109
(cells
mg-')]
=
111
molecules
per
cell
per
s.
Since,
as
we
have
seen,
the
cells
are
making
new
,-lactamases
at
a
slightly
higher
rate,
practically
all
the
P-lactams
slowly
flowing
into
the
periplasm
can
be
trapped
by
this
binding
process,
and
thus
in
theory
the
mechanism
should
be
able
to
produce
a
low
level
of
resistance.
In
practice,
however,
we
have
some
serious
problems.
First,
strain
55M
is
resistant
to
concentrations
of
100
,ug
ml-'
or
more
of
some
of
the
third-generation
antibiotics
(Table
1).
Under
these
conditions,
the
rate
of
influx
of
the
drug
(v)
is
at
least
50
times
higher
than
that
calculated