JCM Accepted Manuscript Posted Online 15 April 2015 J. Clin. Microbiol. doi:10.1128/JCM.00491-15 Copyright © 2015, American Society for Microbiology. All Rights Reserved.
1
Urinary tract infection caused by a capnophilic Proteus mirabilis strain
2 3
Maryse Trapman1, Jakko van Ingen1,2, Jeroen Keijman1, Caroline M. Swanink1
4 5
1. Department of Medical Microbiology, Rijnstate Hospital, Velp, the Netherlands
6
2. Department of Medical Microbiology, Radboud University Medical Center, Nijmegen,
7
the Netherlands
8 9
Corresponding author: Jakko van Ingen, MD, PhD, Department of Medical Microbiology
10
(777), Radboud University Medical Center, PO Box 9101, 6500 HB Nijmegen, the
11
Netherlands. T: +31-24-3614356 ; E: [email protected]
12 13
Word count: 849
14 15
Running Title: Capnophilic Proteus mirabilis
16 17
Summary:
18
From a urine sample of a patient with a urinary tract infection, a carbon dioxide-dependent
19
Proteus mirabilis strain was isolated. It is important to perform urine cultures in 5% carbon
20
dioxide and anaerobic atmosphere if bacteria prominent in Gram stains do not grow on
21
routine media in ambient air.
22
1
23
A 70-year-old male visited his general physician with lower abdominal pain, dysuria and
24
malaise. A urine sample was submitted for staining and culture. The Gram stain showed many
25
erythrocytes, no leukocytes and many Gram-negative rods. A Columbia sheep blood agar
26
(CBA) and MacConkey agar without salt plate (Oxoid, Landsmeer, the Netherlands) were
27
each streaked with 1 microliter of the urine sample and incubated at 35 degrees Celsius in
28
ambient air. After 24 hours, the plates showed no growth. Subsequently, this procedure was
29
repeated, but additionally a Chocolate agar plate was streaked with 1 microliter of urine and
30
incubated at 35 degrees Celsius in a 5% carbon dioxide incubator and a CBA plate was
31
streaked with 1 microliter of urine and incubated at 35 degrees Celsius in an anaerobic jar.
32
The next day, aerobic cultures again showed no growth, but heavy growth of grey waxy
33
colonies was observed on the chocolate agar incubated at 5% carbon dioxide and on the CBA
34
plate incubated in the anaerobic jar. MALDI-TOF mass spectrometry (BioTyper, Bruker
35
Daltonics, Hamburg, Germany) identified the isolate as Proteus mirabilis (score 2.552).
36
To determine the effect of media and environmental conditions on the growth of this strain,
37
we performed subcultures on CBA, chocolate agar and MacConkey agar, in ambient air, in
38
5% carbon dioxide, as well as in anaerobic and microaerophilic jars; this was performed
39
simultaneously for Proteus mirabilis ATCC 7002. The patient’s strain grew on all media at all
40
atmospheric conditions except ambient air. No swarming growth was observed on any
41
medium, at any atmospheric condition. P. mirabilis ATCC 7002 showed swarming growth on
42
all media in all atmospheric conditions. After five subcultures incubated at 5% carbon
43
dioxide, the isolate did not regain capacity to grow in ambient air.
44 45
The isolate was submitted to the National Institute of Public Health and the Environment
46
(RIVM, Bilthoven, the Netherlands) for phenotypic and genotypic identification. The strain
47
was again identified as P. mirabilis, based on 100% sequence similarity of a 500bp fragment
2
48
of the 16S rDNA gene, cell wall lipid content measured by high-performance liquid
49
chromatography and biochemical tests. Growth conditions and lack of swarming growth were
50
completely reproduced.
51
Disk diffusion susceptibility tests were performed according to EUCAST guidelines
52
(available at http:/www.eucast.org), except that plates were incubated in a 5% carbon dioxide
53
incubator. The isolate proved susceptible to all tested compounds except nitrofurantoin (Table
54
1), supporting the identification result. The adjacent nitrofurantoin and norfloxacin disks
55
showed an antagonistic effect of nitrofurantoin on norfloxacin, with a D-shaped zone of
56
inhibition around the norfloxacin disk.
57
Empirically, the patient was treated with oral ciprofloxacin 500mg twice daily for 7 days and
58
reported full resolution of symptoms.
59 60
Failure of aerobic Gram-negative rods to grow in ambient air has been sporadically reported.
61
Given the fact that the reported isolate grew well in carbon dioxide-enriched and anaerobic
62
atmospheres, it is to be considered capnophilic, not microaerophilic. To date, four capnophilic
63
E. coli isolates have been reported as causative agents of human disease; three caused a
64
urinary tract infection (1-3), one a pleural empyema (4). This phenomenon has also been
65
reported for Klebsiella pneumoniae subspecies ozaenae isolates, but without clinical data (5).
66
Recently, a single report of a capnophilic P. mirabilis was published from Japan (6). That
67
isolate, too, was the causative agent of a urinary tract infection and showed growth
68
characteristics similar to the isolate reported here. The growth in microaerophilic atmosphere
69
was not tested for the isolate from Japan and while it did grow in 5% carbon dioxide, growth
70
was less than that of reference strains. Our isolate grew as well as the type strain at 5% carbon
71
dioxide. The isolate reported from Japan showed pin-point colonies without swarming at
72
atmospheric CO2 content of 1 and 2%, with swarming growth at atmospheric CO2 content
3
73
≥3% (6), which we did not observe for our strain. Whether the capnophilicity and lack of
74
swarming result from a single mechanism warrants further investigation.
75
The key role of carbon dioxide for growth of bacteria including Gram-negative facultative
76
aerobic rods, was demonstrated in the 1970s. Repaske and Clayton (7) showed that addition
77
of carbon dioxide to E. coli grown on minimal media with controlled atmospheric condition
78
shortens the lag phase and induces prompt growth. The mechanism behind this effect of
79
carbon dioxide is unknown, as is the mechanism behind the carbon dioxide-dependency in the
80
reported capnophilic isolates (7).
81
Interpretation of antimicrobial susceptibility testing results should be done with caution when
82
performed in 5% carbon dioxide atmosphere, which acidifies the medium. The nitrofurantoin-
83
norfloxacin antagonism is specific to P. mirabilis (8).
84
Routine incubation of urine samples on chocolate medium in a 5% carbon dioxide atmosphere
85
can aid in the recovery of carbon dioxide-dependent organisms, as well as Haemophilus
86
species, which are also capable of causing urinary tract infection (9).
87
In conclusion, the occurrence of carbon dioxide dependent bacteria reinforce the common
88
practice of performing urine cultures in 5% carbon dioxide and anaerobic atmosphere if
89
bacteria prominent in Gram stains of clinical samples do not grow on routine media in
90
ambient air. The mechanisms underlying the capnophilicity and lack of swarming growth
91
warrant further investigation.
92 93 94
References: 1. Tena D, González-Praetorius A, Sáez-Nieto JA, Valdezate S, Bisquert J. 2008.
95
Urinary tract infection caused by capnophilic Escherichia coli. Emerg Infect Dis
96
14:1163-1164.
4
97
2. Essers L, Heinrich WD, Rosenthal E. 1979. Isolation of a carbon dioxide-dependent
98
strain of E. coli from the urine of a patient with chronic pyelonephritis. Zentralbl
99
Bakteriol Orig A 244:229-232. German.
100 101 102
3. Eykyn S, Phillips I. 1978. Carbon dioxide-dependent Escherichia coli. Br Med J
1(6112):576. 4. Lu W, Chang K, Deng S, Li M, Wang J, Xia J, Huang H, Chen M. 2012. Isolation
103
of a capnophilic Escherichia coli strain from an empyemic patient. Diagn Microbiol
104
Infect Dis 73:291-292.
105 106 107
5. Barker J, Brookes G, Johnson T. 1978. Carbon dioxide-dependent klebsiellae. Br
Med J 1(6108):300. 6. Oana K, Yamaguchi M, Nagata M, Washino K, Akahane T, Takamatsu YU,
108
Tsutsui C, Matsumoto T, Kawakami Y. 2013. First isolation of carbon dioxide-
109
dependent Proteus mirabilis from an uncomplicated cystitis patient with Sjögren's
110
syndrome. Jpn J Infect Dis 66:241-244. Erratum in: Jpn J Infect Dis 66:358.
111
7. Repaske R, Clayton MA. 1978. Control of Escherichia coli growth by CO2. J
112 113 114
Bacteriol 135:1162-1164. 8. Shah S, Greenwood D. 1988. Interactions between antibacterial agents of the
quinolone group and nitrofurantoin. J Antimicrob Chemother 21:41-48.
115
9. Dingle TC, Clarridge JE 3rd. 2014. Clinical significance and characterization of
116
Haemophilus influenzae type b genogroup isolates from urine samples in an adult
117
male population. J Clin Microbiol 52:1745-1748.
118 119
5
120
Table 1: Results of disk diffusion drug susceptibility testing Compound
Zone diameter (mm)
Interpretation
Amoxicillin
36
S
Amoxicillin/clavulanic acid
34
S
Cefuroxim
32
S
Co-trimoxazole
36
S
Trimethoprim
22
S
Nitrofurantoin
6
R
Norfloxacin
42*
S
Cirpofloxacin
50
S
Gentamycin
28
S
Fosfomycin
28
S
121
*D-shaped inhibition zone, antagonism with nitrofurantion
122
Tests were performed on Mueller Hinton medium, incubated at 5% carbon dioxide
6
1
Urinary tract infection caused by a capnophilic Proteus mirabilis strain
2 3
Maryse Trapman1, Jakko van Ingen1,2, Jeroen Keijman1, Caroline M. Swanink1
4 5
1. Department of Medical Microbiology, Rijnstate Hospital, Velp, the Netherlands
6
2. Department of Medical Microbiology, Radboud University Medical Center, Nijmegen,
7
the Netherlands
8 9
Corresponding author: Jakko van Ingen, MD, PhD, Department of Medical Microbiology
10
(777), Radboud University Medical Center, PO Box 9101, 6500 HB Nijmegen, the
11
Netherlands. T: +31-24-3614356 ; E: [email protected]
12 13
Word count: 849
14 15
Running Title: Capnophilic Proteus mirabilis
16 17
Summary:
18
From a urine sample of a patient with a urinary tract infection, a carbon dioxide-dependent
19
Proteus mirabilis strain was isolated. It is important to perform urine cultures in 5% carbon
20
dioxide and anaerobic atmosphere if bacteria prominent in Gram stains do not grow on
21
routine media in ambient air.
22
1
23
A 70-year-old male visited his general physician with lower abdominal pain, dysuria and
24
malaise. A urine sample was submitted for staining and culture. The Gram stain showed many
25
erythrocytes, no leukocytes and many Gram-negative rods. A Columbia sheep blood agar
26
(CBA) and MacConkey agar without salt plate (Oxoid, Landsmeer, the Netherlands) were
27
each streaked with 1 microliter of the urine sample and incubated at 35 degrees Celsius in
28
ambient air. After 24 hours, the plates showed no growth. Subsequently, this procedure was
29
repeated, but additionally a Chocolate agar plate was streaked with 1 microliter of urine and
30
incubated at 35 degrees Celsius in a 5% carbon dioxide incubator and a CBA plate was
31
streaked with 1 microliter of urine and incubated at 35 degrees Celsius in an anaerobic jar.
32
The next day, aerobic cultures again showed no growth, but heavy growth of grey waxy
33
colonies was observed on the chocolate agar incubated at 5% carbon dioxide and on the CBA
34
plate incubated in the anaerobic jar. MALDI-TOF mass spectrometry (BioTyper, Bruker
35
Daltonics, Hamburg, Germany) identified the isolate as Proteus mirabilis (score 2.552).
36
To determine the effect of media and environmental conditions on the growth of this strain,
37
we performed subcultures on CBA, chocolate agar and MacConkey agar, in ambient air, in
38
5% carbon dioxide, as well as in anaerobic and microaerophilic jars; this was performed
39
simultaneously for Proteus mirabilis ATCC 7002. The patient’s strain grew on all media at all
40
atmospheric conditions except ambient air. No swarming growth was observed on any
41
medium, at any atmospheric condition. P. mirabilis ATCC 7002 showed swarming growth on
42
all media in all atmospheric conditions. After five subcultures incubated at 5% carbon
43
dioxide, the isolate did not regain capacity to grow in ambient air.
44 45
The isolate was submitted to the National Institute of Public Health and the Environment
46
(RIVM, Bilthoven, the Netherlands) for phenotypic and genotypic identification. The strain
47
was again identified as P. mirabilis, based on 100% sequence similarity of a 500bp fragment
2
48
of the 16S rDNA gene, cell wall lipid content measured by high-performance liquid
49
chromatography and biochemical tests. Growth conditions and lack of swarming growth were
50
completely reproduced.
51
Disk diffusion susceptibility tests were performed according to EUCAST guidelines
52
(available at http:/www.eucast.org), except that plates were incubated in a 5% carbon dioxide
53
incubator. The isolate proved susceptible to all tested compounds except nitrofurantoin (Table
54
1), supporting the identification result. The adjacent nitrofurantoin and norfloxacin disks
55
showed an antagonistic effect of nitrofurantoin on norfloxacin, with a D-shaped zone of
56
inhibition around the norfloxacin disk.
57
Empirically, the patient was treated with oral ciprofloxacin 500mg twice daily for 7 days and
58
reported full resolution of symptoms.
59 60
Failure of aerobic Gram-negative rods to grow in ambient air has been sporadically reported.
61
Given the fact that the reported isolate grew well in carbon dioxide-enriched and anaerobic
62
atmospheres, it is to be considered capnophilic, not microaerophilic. To date, four capnophilic
63
E. coli isolates have been reported as causative agents of human disease; three caused a
64
urinary tract infection (1-3), one a pleural empyema (4). This phenomenon has also been
65
reported for Klebsiella pneumoniae subspecies ozaenae isolates, but without clinical data (5).
66
Recently, a single report of a capnophilic P. mirabilis was published from Japan (6). That
67
isolate, too, was the causative agent of a urinary tract infection and showed growth
68
characteristics similar to the isolate reported here. The growth in microaerophilic atmosphere
69
was not tested for the isolate from Japan and while it did grow in 5% carbon dioxide, growth
70
was less than that of reference strains. Our isolate grew as well as the type strain at 5% carbon
71
dioxide. The isolate reported from Japan showed pin-point colonies without swarming at
72
atmospheric CO2 content of 1 and 2%, with swarming growth at atmospheric CO2 content
3
73
≥3% (6), which we did not observe for our strain. Whether the capnophilicity and lack of
74
swarming result from a single mechanism warrants further investigation.
75
The key role of carbon dioxide for growth of bacteria including Gram-negative facultative
76
aerobic rods, was demonstrated in the 1970s. Repaske and Clayton (7) showed that addition
77
of carbon dioxide to E. coli grown on minimal media with controlled atmospheric condition
78
shortens the lag phase and induces prompt growth. The mechanism behind this effect of
79
carbon dioxide is unknown, as is the mechanism behind the carbon dioxide-dependency in the
80
reported capnophilic isolates (7).
81
Interpretation of antimicrobial susceptibility testing results should be done with caution when
82
performed in 5% carbon dioxide atmosphere, which acidifies the medium. The nitrofurantoin-
83
norfloxacin antagonism is specific to P. mirabilis (8).
84
Routine incubation of urine samples on chocolate medium in a 5% carbon dioxide atmosphere
85
can aid in the recovery of carbon dioxide-dependent organisms, as well as Haemophilus
86
species, which are also capable of causing urinary tract infection (9).
87
In conclusion, the occurrence of carbon dioxide dependent bacteria reinforce the common
88
practice of performing urine cultures in 5% carbon dioxide and anaerobic atmosphere if
89
bacteria prominent in Gram stains of clinical samples do not grow on routine media in
90
ambient air. The mechanisms underlying the capnophilicity and lack of swarming growth
91
warrant further investigation.
92 93 94
References: 1. Tena D, González-Praetorius A, Sáez-Nieto JA, Valdezate S, Bisquert J. 2008.
95
Urinary tract infection caused by capnophilic Escherichia coli. Emerg Infect Dis
96
14:1163-1164.
4
97
2. Essers L, Heinrich WD, Rosenthal E. 1979. Isolation of a carbon dioxide-dependent
98
strain of E. coli from the urine of a patient with chronic pyelonephritis. Zentralbl
99
Bakteriol Orig A 244:229-232. German.
100 101 102
3. Eykyn S, Phillips I. 1978. Carbon dioxide-dependent Escherichia coli. Br Med J
1(6112):576. 4. Lu W, Chang K, Deng S, Li M, Wang J, Xia J, Huang H, Chen M. 2012. Isolation
103
of a capnophilic Escherichia coli strain from an empyemic patient. Diagn Microbiol
104
Infect Dis 73:291-292.
105 106 107
5. Barker J, Brookes G, Johnson T. 1978. Carbon dioxide-dependent klebsiellae. Br
Med J 1(6108):300. 6. Oana K, Yamaguchi M, Nagata M, Washino K, Akahane T, Takamatsu YU,
108
Tsutsui C, Matsumoto T, Kawakami Y. 2013. First isolation of carbon dioxide-
109
dependent Proteus mirabilis from an uncomplicated cystitis patient with Sjögren's
110
syndrome. Jpn J Infect Dis 66:241-244. Erratum in: Jpn J Infect Dis 66:358.
111
7. Repaske R, Clayton MA. 1978. Control of Escherichia coli growth by CO2. J
112 113 114
Bacteriol 135:1162-1164. 8. Shah S, Greenwood D. 1988. Interactions between antibacterial agents of the
quinolone group and nitrofurantoin. J Antimicrob Chemother 21:41-48.
115
9. Dingle TC, Clarridge JE 3rd. 2014. Clinical significance and characterization of
116
Haemophilus influenzae type b genogroup isolates from urine samples in an adult
117
male population. J Clin Microbiol 52:1745-1748.
118 119
5
120
Table 1: Results of disk diffusion drug susceptibility testing Compound
Zone diameter (mm)
Interpretation
Amoxicillin
36
S
Amoxicillin/clavulanic acid
34
S
Cefuroxim
32
S
Co-trimoxazole
36
S
Trimethoprim
22
S
Nitrofurantoin
6
R
Norfloxacin
42*
S
Cirpofloxacin
50
S
Gentamycin
28
S
Fosfomycin
28
S
121
*D-shaped inhibition zone, antagonism with nitrofurantion
122
Tests were performed on Mueller Hinton medium, incubated at 5% carbon dioxide
6
