J Antimicrob Chemother 2015; 70: 3267 – 3272
doi:10.1093/jac/dkv251 Advance Access publication 25 August 2015
Gradual increase in antibiotic concentration affects persistence
of Klebsiella pneumoniae
Huan Ren1†, Xin He1†, Xiaoli Zou1, Guoqing Wang1, Shuhua Li2 and Yanxia Wu1*
1
Department of Public Health Laboratory Sciences, West China School of Public Health, Sichuan University, Chengdu, Sichuan, China;
2
No.4 West China Teaching Hospital, West China School of Public Health, Sichuan University, Chengdu, Sichuan, China
*Corresponding author. Tel: +86-28-85502097; E-mail: [email protected]
†These authors contributed equally to this study.
Received 17 February 2015; returned 19 April 2015; revised 1 June 2015; accepted 21 July 2015
Objectives: Sublethal bactericidal antibiotics promote the formation of multidrug-tolerant persisters. Clinically,
serum drug concentration increases gradually and reaches the peak level with high killing efficiency some time
after administration. This study aimed to investigate if the initial low antibiotic concentration would promote
persister formation in Klebsiella pneumoniae, an increasingly important nosocomial pathogen.
Methods: Time-dependent killings of K. pneumoniae by different types of bactericidal antibiotics were conducted
to determine the existence of multidrug-tolerant K. pneumoniae persisters. Killing experiments with antibiotic
gradient feeding were then conducted for a K. pneumoniae laboratory strain (ATCC 10031) and a clinical isolate
(YWSCU-03) by adding antibiotics step by step until the drug peak serum concentration was attained.
Results: Multidrug-tolerant persisters indeed existed in K. pneumoniae and the persistence decreased with
increasing drug concentrations or prolonged treatments. Antibiotic gradient feeding, to simulate a gradual
increase in serum drug concentration, not only significantly elevated the persistence of ATCC 10031 and
YWSCU-03, but also increased the frequency of drug-resistant mutant formation in YWSCU-03.
Conclusions: After administration, the initial low serum drug concentration could promote the formation of
multidrug-tolerant bacterial persisters, which could survive the lethal drug concentrations attained later and
potentially render the antibiotic treatment fruitless. Therefore, antibiotic treatments should be based on the
comprehensive analysis of, not only drug pharmacokinetics, but also the synergistic effect between pharmacokinetics and persister formation.
Introduction
Persisters are a subpopulation of susceptible bacteria that survive
lethal doses of antibiotics due to their inactive physiological state.
Distinct from drug-resistant mutants with genetic modification, persisters do not replicate in the presence of antibiotics.
However, once the drugs are removed, they resume growth and
form a population equally susceptible to the antibiotics as the
parental population. Due to such a transient trait of multidrug tolerance, they play an important role in the relapse of microbial
infections, especially those associated with biofilms, which usually harbour a high level of persisters.1,2 Until now, persister cells
have been described in many microbial species, but never in
Klebsiella pneumoniae, which is an important nosocomial pathogen
gaining growing clinical concerns and frequently found in
biofilm-associated chronic infections, both invasive device and
non-device related. The use of antibiotics is frequently seen to be
insufficient to eradicate infections by K. pneumoniae. This is usually
attributed to the development of drug resistance and the possible
role of K. pneumoniae persisters has never been reported.
Increasing evidence suggests that various intracellular stress
responses, such as SOS response,3 – 5 oxidative stress response6,7
and stringent response,8 can turn some cells in a bacterial population into multidrug-tolerant persisters. Pretreatment with
fluoroquinolones around the MIC elevates the persistence of
Escherichia coli by .1000-fold through a complicated process involving the repair of DNA double-strand breaks (DSBs), collapsed
replication forks, stalled transcription complexes3 – 5 and the dissipation of proton motive force (PMF).4 In E. coli, oxidative stress
response governed by OxyR and SoxRS regulons, which control
the responses to hydrogen peroxide and superoxide, respectively,
is also involved in the formation of persisters.6,7,9 Bactericidal antibiotics generate reactive oxygen species (ROS) including superoxide, hydrogen peroxide and hydroxyl radicals in bacterial cells,
which damage macromolecules such as DNA, lipids and proteins,
subsequently inducing bacterial stress responses including SOS
# The Author 2015. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
For Permissions, please e-mail: [email protected]
3267
Ren et al.
response and oxidative stress response.10,11 In K. pneumoniae,
sub-MIC aminoglycosides significantly generate ROS and induce
SOS response,12 while paraquat and hydrogen peroxide induce
oxidative stress response regulated by RpoS, SoxRS and YjcC.13
These findings together suggest that antibiotic treatments may
induce the formation of persisters in K. pneumoniae. ROS generation is usually stronger at relatively low antibiotic concentrations.14 Bactericidal antibiotics far below clinical concentrations
significantly up-regulate oxyR and soxRS.11 Fluoroquinolones
and hydrogen peroxide at sublethal concentrations promote persister formation more efficiently.3,7 During clinical antibiotic treatment, drug serum concentration increases gradually and peaks
some time after administration. It would therefore be particularly
interesting and clinically relevant to investigate if the initial low
concentration of antibiotic would promote the formation of
K. pneumoniae persisters, which would survive the lethal concentration later and consequently reduce the efficacy of treatments.
For this purpose, this study utilizes antibiotic gradient feeding for a
laboratory strain and a clinical strain of K. pneumoniae, where
antibiotic concentration was increased step by step until the
drug serum peak concentration was reached and maintained
in order to test if persisters induced by the initial low antibiotic
concentration would endure the lethal concentrations.
sterile 1% NaCl. The suspension was then serially diluted 10-fold and spotted onto LB agar for colony counting. The detection limit was 10 cfu/mL.
Persistence was evaluated by the number of surviving cells on the plateau
of killing curves.
For the killing experiments with antibiotic gradient feeding, antibiotics
were added step by step until the maximum concentration was reached at
2 h, as listed in Table 1. The parameters of antibiotic gradient feeding were
determined according to the pharmacokinetics of specific antibiotics
(Table 2). Killings with the initial low concentration and the final maximum
concentration were conducted simultaneously as controls.
Inheritability of drug tolerance
An exponentially growing population of ATCC 10031 was challenged with
4 mg/L ciprofloxacin (200× MIC) for 3 h, 50 mg/L kanamycin (33× MIC) for
6 h or 5 mg/L ceftriaxone (100× MIC) for 6 h. At designated timepoints,
samples were withdrawn for cfu counting. The survivors at the end of
treatment were washed three times and then cultured in fresh LB medium
overnight. The overnight culture was then diluted 1: 1000 in LB broth and
grown to exponential phase when the population was challenged again as
described above. The cycle was repeated for three consecutive days.
Results
Persisters exist in the K. pneumoniae population
Materials and methods
Bacterial strains and antibiotics
K. pneumoniae laboratory strain ATCC 10031 and a recently isolated clinical strain YWSCU-03 were used in this study. LB broth was used throughout the study except that Mueller – Hinton broth was used for MIC
measurement. The antibiotics used in this study were ciprofloxacin
(Sigma – Aldrich, USA), kanamycin (Amresco, USA) and ceftriaxone
(Sichuan Pharmaceutical Preparation Co., China). The purity of ceftriaxone
was measured to be 82.6% by HPLC and ceftriaxone concentration was
adjusted accordingly in all related experiments.
Measurement of persistence
Persistence of K. pneumoniae was measured by time-dependent killing
experiments. An overnight culture was diluted 1 : 1000 in LB broth and
incubated at 378C (180 rpm) for 3 h. The culture was then divided into aliquots for a 24 h challenge by ciprofloxacin, kanamycin or ceftriaxone at
different concentrations, with three replicates for each condition. At designated timepoints of antibiotic treatment, a 0.5 mL sample from each aliquot was withdrawn and then washed twice with and resuspended in
Ciprofloxacin, kanamycin and ceftriaxone are commonly used
for clinical treatment of infections caused by K. pneumoniae.
Ciprofloxacin is a fluoroquinolone damaging DNA. Kanamycin is
an aminoglycoside inhibiting protein synthesis. Ceftriaxone is a
b-lactam antibiotic that disrupts synthesis of the peptidoglycan
layer of the bacterial cell wall. The time-dependent killing curves
of the three antibiotics at different concentrations not only
displayed antibiotic-specific characteristics, but also shared
common patterns.
The killing curves all presented the typical biphasic pattern: a
rapid killing phase was followed by a slow phase, which is the
plateau representing the level of persistence. The only exception
was the killing by 0.1 mg/L ciprofloxacin, where the number of
survivors decreased in the first 5 h and then elevated quickly
(Figure 1a), indicating the generation of drug-resistant mutants
that can replicate in the presence of ciprofloxacin. Given ciprofloxacin mean serum concentrations 12 h after dosing with 250,
500 or 750 mg are 0.1, 0.2 and 0.4 mg/L, respectively,15 this result
suggests low ciprofloxacin serum concentrations between doses
may accelerate the development of bacterial drug resistance.
Table 1. Parameters for the gradient feeding of antibiotics
Drug concentration (mg/L)
Antibiotic
0h
0.5 h
1.0 h
1.5 h
2.0 h
MIC (mg/L)
0.5
0.75
2.0
3.0
4.0
5.0
0.02
0.03
Ciprofloxacin
ATCC 10031
YWSCU-03
0.02
0.03
0.2
0.3
Kanamycin
ATCC 10031
YWSCU-03
2.0
4.0
3.0
6.0
8.0
15.0
16.0
28.0
20.0
35.0
1.5
3.0
Ceftriaxone
ATCC 10031
YWSCU-03
0.1
0.1
1.0
1.0
5.0
5.0
25.0
25.0
80.0
80.0
0.05
0.06
3268
JAC
Antibiotic pharmacokinetics and bacterial persistence
Table 2. Pharmacokinetics of bactericidal antibiotics used in this study
Antibiotic
Ciprofloxacin
Kanamycin
Ceftriaxone
Administration
Peak serum concentration (mg/L)
Peak time (h)
Usage
Reference
oral
muscle injection
muscle injection
4
20
80
2
1 –2
2 –3
0.75 g/12 h
0.5 g/12 h
1.0 g/24 h
15
16,18
19
10
9
8
7
6
5
4
3
2
1
0
Log10 cfu/mL
Log10 cfu/mL
CIP 0.1
CIP 0.25
CIP 0.5
CIP 1.0
10
KAN 15.0
KAN 50.0
8
6
4
2
0
4
8
12
16
Time (h)
20
0
24
0
4
8
0
4
8
12
16
Time (h)
20
24
2
3
Time (h)
9
8
7
6
5
4
3
2
4
24
Day 1
Day 2
Day 3
0
1
2
3
Ceftriaxone
(f)
Day 1
Day 2
Day 3
1
20
Time (h)
Kanamycin
0
16
12
Time (h)
Ciprofloxacin
(d)
CRO 0.5
CRO 1.0
CRO 2.5
CRO 5.0
(e)
9
8
7
6
5
4
3
2
1
0
KAN 5.0
KAN 25.0
KAN 100.0
12
Log10 cfu/mL
Log10 cfu/mL
(c)
(b)
10
9
8
7
6
5
4
3
2
5
6
Log10 cfu/mL
Log10 cfu/mL
(a)
9
8
7
6
5
4
3
2
1
0
Day 1
Day 2
Day 3
0
2
4
6
Time (h)
Figure 1. Killing curves of K. pneumoniae. (a) Ciprofloxacin at 0.1, 0.25, 0.5 and 1.0 mg/L. (b) Kanamycin at 5, 15, 25, 50 and 100 mg/L. (c) Ceftriaxone at
0.5, 1.0, 2.5 and 5.0 mg/L. CIP, ciprofloxacin; KAN, kanamycin; CRO, ceftriaxone. From the results of (a) – (c), the inheritability of the tolerance of
K. pneumoniae was tested with (d) 4.0 mg/L ciprofloxacin, (e) 50.0 mg/L kanamycin and (f) 5.0 mg/L ceftriaxone.
Besides the biphasic pattern, the persistence to all three
antibiotics decreased with increasing drug concentration.
Persistence to ciprofloxacin decreased from 106 to 104 cfu/mL
with an increase in the ciprofloxacin concentration from 0.25
to 1.0 mg/L (Figure 1a), presumably because the level of DSBs
went beyond the rescuing capability of the SOS response.
Interestingly, the plateau of killing curves gradually collapsed
between 5 and 24 h and converged to 10 3 – 10 4 cfu/mL by
24 h, indicating that the persister population consisted of
both ‘fragile’ and ‘robust’ persister cells. Persistence to kanamycin decreased from 106 to 10 – 100 cfu/mL when the kanamycin concentration increased from 5 to 50 mg/L (Figure 1b).
Persistence to ceftriaxone at 24 h decreased from 104 cfu/mL
to a level below the detection limit (10 cfu/mL) when the
3269
Ren et al.
ceftriaxone concentration increased from 0.5 to 5.0 mg/L
(Figure 1c). Clinically, ceftriaxone serum concentration is usually .5 mg/L within 24 h after injection, 19 thus this result
implies that the clinical persistence of K. pneumoniae ATCC
10031 to ceftriaxone could be very low.
The inheritability of the multidrug tolerance of persisters illustrated in Figure 1(a –c) was tested and the results are shown in
Figure 1(d – f). Persister cells that survived the treatment by high
concentration antibiotics on day 1 were regrown into a population
on day 2, of which the majority of bacterial cells were killed rapidly
by the same concentration of antibiotics and again a similar level
of persisters survived. The same pattern repeated on day 3. This
result demonstrated that the multidrug tolerance of these survivors was transient and not inheritable, confirming these cells as
persisters, in contrast to drug-resistant mutants that carry inheritable drug resistance mutations.
0
4
8
12
16
Time (h)
20
CIP 0.03
CIP gradient
CIP 5.0
6
4
0
24
0
4
8
(d)
12
16
Time (h)
20
24
YWSCU–03
12
10
KAN 2.0
KAN gradient
KAN 20.0
Log10 cfu/mL
Log10 cfu/mL
8
2
8
6
KAN 4.0
KAN gradient
KAN 35.0
4
2
0
4
8
(e)
12
16
Time (h)
20
0
24
(f)
ATCC 10031
9
8
7
6
5
4
3
2
1
0
Log10 cfu/mL
CIP 0.02
CIP gradient
CIP 4.0
ATCC 10031
10
9
8
7
6
5
4
3
2
1
0
YWSCU–03
10
(c)
Log10 cfu/mL
(b)
ATCC 10031
10
9
8
7
6
5
4
3
2
Clinically, it takes hours for the serum drug concentration to reach
a lethal level, especially when administered orally or intramuscularly. To test if the initial low antibiotic concentration would elevate bacterial persistence for the reasons stated earlier, we
conducted time-dependent killing experiments with gradient
feeding of antibiotics, with the parameters listed in Table 1. In
contrast to the clinical setting where the drug concentration
decreases gradually after the peak time, in this study the drug
concentration was maintained at the peak level afterwards to
test if persisters resulting from the initial low drug concentrations
could survive lethal antibiotic concentrations.
As shown in Figure 2(a), ATCC 10031 culture challenged by
ciprofloxacin at the MIC (the low concentration control) grew to
0
4
8
12
16
Time (h)
20
24
YWSCU–03
9
8
CRO 0.1
CRO gradient
CRO 80.0
Log10 cfu/mL
Log10 cfu/mL
(a)
Gradient feeding of different types of bactericidal
antibiotics elevates the persistence of K. pneumoniae
7
CRO 0.1
CRO gradient
CRO 80.0
6
5
4
3
0
4
8
12
16
Time (h)
20
24
2
0
4
8
12
16
Time (h)
20
24
Figure 2. Effect of gradient feeding of antibiotics on the persistence of K. pneumoniae laboratory strain ATCC 10031 and clinical strain YWSCU-03. (a)
ATCC 10031, ciprofloxacin. (b) YWSCU-03, ciprofloxacin. (c) ATCC 10031, kanamycin. (d) YWSCU-03, kanamycin. (e) ATCC 10031, ceftriaxone. (f)
YWSCU-03, ceftriaxone. Filled diamonds, antibiotic low concentration control for ATCC 10031; opens diamonds, antibiotic low concentration control
for YWSCU-03; filled circles, gradient antibiotic feeding for ATCC 10031; open circles, gradient antibiotic feeding for YWSCU-03; crosses, antibiotic
high concentration control. CIP, ciprofloxacin; KAN, kanamycin; CRO, ceftriaxone.
3270
JAC
Antibiotic pharmacokinetics and bacterial persistence
stationary phase after a slight killing phase, demonstrating an
adaptive process with the generation of drug-resistant mutants.
In contrast, 4.0 mg/L ciprofloxacin (the high concentration control) killed the bacteria efficiently and the number of survivors
reached the plateau of 3 log that lasted until the end of the treatment. During the challenge with antibiotic gradient feeding, the
initial low drug concentration enabled some cells to adapt and
resulted in 10-fold more persisters compared with the killing
by 4.0 mg/L ciprofloxacin alone. The same experiment was
conducted for the clinical isolate YWSCU-03. The persistence,
however, could not be determined since drug-resistant mutants
arose under all killing conditions in this experiment (Figure S1a,
available as Supplementary data at JAC Online). The peak serum
concentration of ciprofloxacin at a 1000 mg oral dose is 5.4 mg/L;15
therefore, we set the maximum concentration at 5.0 mg/L and
adjusted the parameters for ciprofloxacin gradient feeding
accordingly as in Table 1. The results shown in Figure 2(b) confirmed that ciprofloxacin gradient feeding similarly elevated the
persistence of YWSCU-03 by 10-fold and the persisters endured
the peak concentration for ≥22 h.
Similarly, in comparison with 20.0 mg/L kanamycin, the gradient feeding of kanamycin resulted in .10-fold more survivors in
ATCC 10031 during the treatment (Figure 2c). Kanamycin is usually administrated intramuscularly every 12 h. For 20.0 mg/L
kanamycin, there were 100 cfu/mL persisters by 12 h and nearly
no survivors by 24 h. For kanamycin gradient feeding, the persistence was .1000 cfu/mL by 12 h and .10 cfu/mL by 24 h. The
same experiment conducted for YWSCU-03 exhibited 2 log
more persisters to the killing by gradient feeding than the killing
by 20.0 mg/L kanamycin by 6 h, while drug-resistant mutants
were generated during both killings (Figure S1b). Due to the potential ototoxicity and nephrotoxicity of kanamycin, clinically the concentrations prior to the next dose should not exceed 5 –10 mg/L,
desirable kanamycin serum peak concentrations are generally
15 – 30 mg/L and prolonged kanamycin peak concentrations
.35 mg/L should be avoided.16 We therefore increased the maximum concentration from 20.0 to 35.0 mg/L and adjusted the
concentration at each timepoint accordingly as listed in Table 1
for the experiment with YWSCU-03. Persistence to 35.0 mg/L
kanamycin was steady at 100 cfu/mL after 6 h. The number of
survivors to kanamycin gradient feeding was higher than that to
35.0 mg/L kanamycin before 6 h, but was hard to evaluate after
6 h owing to the emergence of drug-resistant cells (Figure 2d).
The peak serum concentration of ceftriaxone (80.0 mg/L) was
.1000-fold higher than the MICs for ATCC 10031 and YWSCU-03.
Ceftriaxone at such an unusually high concentration reduced
the number of ATCC 10031 survivors below the detection limit by
6 h (Figure 2e), but .1000 cfu/mLYWSCU-03 cells persisted through
the treatment (Figure 2f). For ceftriaxone gradient feeding, the persistence of ATCC 10031 was ,100 cfu/mL at 6 h and decreased to
the detection limit by 24 h. Persistence of YWSCU-03 to ceftriaxone
gradient feeding was .104 cfu/mL at 6 h, after which it was hard
to evaluate due to the replicating drug-resistant mutants.
Taken together, in comparison with the high concentration
control, antibiotic gradient feeding elevated the persistence of
the laboratory strain and the clinical isolate of K. pneumoniae
used in this study and tended to increase the frequency of drug
resistance mutations in the clinical strain, indicating that clinically
a gradual increase in serum drug concentration may reduce the
efficacy of antibiotic treatment.
Discussion
In this work, we demonstrated the existence of persisters in populations of K. pneumoniae laboratory strain ATCC 10031 and clinical
isolate YWSCU-03, characterized the persistence by the killing
curves to different bactericidal antibiotics commonly used for clinical treatment of infections by K. pneumoniae and further examined the effect of the gradual increase in drug concentration on
the persistence of both strains.
Overall, the persistence varies with drug concentration,
treatment duration and specific antibiotics. The persistence of
K. pneumoniae decreased with increasing antibiotic concentrations and prolonged antibiotic treatment. This phenomenon is
not unique to K. pneumoniae, but ubiquitous to many types of
bacteria. Fluoroquinolone antibiotics induce the formation of
E. coli persister cells through LexA-regulated SOS response in
a concentration-dependent manner.3,6 Induction of the SOS
response up-regulates the expression of DNA repair-related proteins to rescue cells and promote persister formation. TisB, a peptide toxin induced upon strong SOS induction, forms pores in the
outer membrane of E. coli and turns the cell into deeply inactive
persisters by depleting PMF.3,4 However, with increasing fluoroquinolone concentration, more cells die due to DNA damage beyond their
repair capability. With prolonged fluoroquinolone treatment, persisters that lost the balance between DNA damage and DNA repair will
gradually die. Persisters formed through the TisB pathway would be
robust since they are in a deeply inactive state. Although the SOS
response and DNA repair are widely conserved in various species,
the pathway of SOS in K. pneumoniae has not been elucidated so
far. The mechanism of persister formation through SOS response
described above was elucidated in E. coli. If it is similar in K. pneumoniae, it would be reasonable to speculate that DNA repair played
the key role in the concentration-dependent manner of K. pneumoniae persistence to ciprofloxacin (as shown in Figure 1a), and that
the converged persistence at 24 h in Figure 1(a) and the steady-flat
plateaus of killing curves by high concentration ciprofloxacin in
Figure 2(a, b) represented the robust and deeply inactive K. pneumoniae persister cells since they were not affected by drug concentrations and treatment durations.
Until now, the microbial killing mechanisms of aminoglycosides and b-lactams have not been reported to be involved in
persister formation, but these antibiotics can generate ROS that
induce the bacterial SOS response and oxidative stress response.
Oxidative stress response induces persister formation with the
involvement of SOS response;6 therefore, ROS may promote persister formation ultimately through DNA repair. This is consistent
with our observation that persistence to kanamycin or ceftriaxone
gradually decreased during prolonged treatment or with increasing drug concentrations.
For the experiments with antibiotic gradient feeding in this
study, antibiotic concentration was increased step by step to
attain the serum peak concentration and maintained at the
peak value afterwards. The number of ATCC 100131 persisters
to antibiotic gradient feeding gradually decreased to a low level
at the end of the 24 h treatment. However, clinically the number
of survivors to one dose of antibiotics should be much higher since
drug serum concentration will decrease after peak time. For
example, the interval of drug administration is 12 h for ciprofloxacin and kanamycin and their elimination half-life is 6 h. The persistence of ATCC 10031 prior to the next dose would therefore be
3271
Ren et al.
.104 cfu/mL for ciprofloxacin and .103 cfu/mL for kanamycin
(Figure 2a and c).
Compared with the laboratory strain, the negative effect
of a gradual increase in antibiotic concentration on clinical
strains would be more significant. In this study, the number of
YWSCU-03 survivors was significantly higher than ATCC 10031 in
all cases under the same killing conditions. The antibiotic gradient
feeding not only significantly promoted persister formation in
YWSCU-03, but obviously increased the frequency of drug resistance mutations. Sublethal concentrations of bactericidal antibiotics could induce mutagenesis and cause heterogeneous
increases in the MICs of a range of antibiotics through ROS generation.17 Repeated antibiotic treatment of patients has been
reported to select high-persister (hip) mutants.2 Therefore, the initial low serum antibiotic concentration could potentially facilitate
the development of hip mutants and drug-resistant mutants, and
the results of YWSCU-03 experiments described above may serve
as strong evidence for this.
In clinical reality, the host environment is much more complicated than this in vitro study. For example, the drug concentration
in infection sites might be much lower than serum drug concentrations, the killing power of antibiotics might be compromised by
various host substances and local physical factors such as pH. In
addition, the clinical pathogens are usually much less susceptible
than the strains used in this study. Therefore, the actual effect of
such a gradual increase in drug serum concentration on bacterial
persistence would potentially be more significant. The optimization of antibiotic therapeutics should be based not only on the
pharmacokinetics of drugs, but also on the synergistic effect of
pharmacokinetics and bacterial persistence as found in this study.
Funding
This work was supported by the Starting Research Fund to Talents from
Overseas to Y. W. provided by Sichuan University (China) (Grant No.
2082204184001; Project No. YJ201350) and the Science Foundation to
Returned Overseas Chinese Scholars provided by the State Education
Ministry of China (Grant No. 2014-1685-11-8).
Transparency declarations
None to declare.
Supplementary data
Figure S1 is available as Supplementary data at JAC Online (http://jac.
oxfordjournals.org/).
3272
References
1 Lewis K. Persister cells and the riddle of biofilm survival. Biochemistry
(Mosc) 2005; 70: 267– 74.
2 Mulcahy LR, Burns JL, Lory S et al. Emergence of Pseudomonas aeruginosa strains producing high levels of persister cells in patients with cystic
fibrosis. J Bacteriol 2010; 192: 6191 –9.
3 Dorr T, Lewis K, Vulic M. SOS response induces persistence to fluoroquinolones in Escherichia coli. PLoS Genet 2009; 5: e1000760.
4 Dorr T, Vulic M, Lewis K. Ciprofloxacin causes persister formation by
inducing the TisB toxin in Escherichia coli. PLoS Biol 2010; 8: e1000317.
5 Theodore A, Lewis K, Vulic M. Tolerance of Escherichia coli to fluoroquinolone antibiotics depends on specific components of the SOS response
pathway. Genetics 2013; 195: 1265– 76.
6 Wu Y, Vulic M, Keren I et al. Role of oxidative stress in persister tolerance.
Antimicrob Agents Chemother 2012; 56: 4922– 6.
7 Vega NM, Allison KR, Khalil AS et al. Signaling-mediated bacterial
persister formation. Nat Chem Biol 2012; 8: 431–3.
8 Amato SM, Brynildsen MP. Nutrient transitions are a source of persisters
in Escherichia coli biofilms. PLoS One 2014; 9: e93110.
9 Grant SS, Hung DT. Persistent bacterial infections, antibiotic tolerance,
and the oxidative stress response. Virulence 2013; 4: 273–83.
10 Kohanski MA, Dwyer DJ, Hayete B et al. A common mechanism
of cellular death induced by bactericidal antibiotics. Cell 2007; 130:
797–810.
11 Dwyer DJ, Belenky PA, Yang J et al. Antibiotics induce redox-related
physiological alterations as part of their lethality. Proc Natl Acad Sci USA
2014; 111: E2100– 9.
12 Baharoglu Z, Krin E, Mazel D. RpoS plays a central role in the SOS induction by sub-lethal aminoglycoside concentrations in Vibrio cholerae. PLoS
Genet 2013; 9: e1003421.
13 Huang CJ, Wang ZC, Huang HY et al. YjcC, a c-di-GMP phosphodiesterase protein, regulates the oxidative stress response and virulence of
Klebsiella pneumoniae CG43. PLoS One 2013; 8: e66740.
14 Keren I, Wu Y, Inocencio J et al. Killing by bactericidal antibiotics does
not depend on reactive oxygen species. Science 2013; 339: 1213 –6.
15 Bayer HealthCare Pharmaceuticals. Information on Ciprofloxacin
Pharmacokinetics: CIPROw Ciprofloxacin Hydrochloride. http://www.
accessdata.fda.gov/drugsatfda_docs/label/2013/019537s082,02078
0s040lbl.pdf.
16 Information on Kanamycin Pharmacokinetics [in English]. http://www.
drugs.com/monograph/kanamycin-sulfate.html.
17 Kohanski MA, DePristo MA, Collins JJ. Sublethal antibiotic treatment
leads to multidrug resistance via radical-induced mutagenesis. Mol Cell
2010; 37: 311–20.
18 Information on Kanamycin Pharmacokinetics [in Chinese]. http://www.
sosoyy.com/product/product-105550.html.
19 Genentech USA. Ceftriaxone Pharmacokinetics Information: Rochephin_
Prescribing. Revised: April 2013. http://www.gene.com/download/pdf/
rocephin_prescribing.pdf.