Article

Brucella abortus bacA mutant induces greater pro-inflammatory cytokines than the wild-type parent strain

Department of Veterinary and Animal Sciences, University of Massachusetts, Amherst, MA 01003, USA.
Microbes and Infection (Impact Factor: 2.86). 02/2007; 9(1):55-62. DOI: 10.1016/j.micinf.2006.10.008
Source: PubMed

ABSTRACT

The inner-membrane protein BacA affects Brucella LPS structure. A bacA deletion mutant of Brucella abortus, known as KL7 (bacA(mut)-KL7), is attenuated in BALB/c mice and protects against challenge. Thus, bacA mutation was a candidate for incorporation into live attenuated vaccines. We assessed bacA(mut)-KL7 in 2 additional mouse strains: the more resistant C57BL/6 that produces interferon-gamma throughout the infection and the highly susceptible interferon-gamma-deficient C57BL/6 in which brucellae exhibit continual exponential growth. While it was hypothesized that bacA(mut)-KL7 would exhibit even greater attenuation relative to its parent strain B. abortus 2308 in C57BL/6 mice than it did in BALB/c mice, this was not the case. Moreover, it was more pathogenic in C57BL/6 interferon-gamma-deficient mice than 2308 causing abscesses and wasting even though the splenic loads of bacA(mut)-KL7 were significantly lower. These 2 observations were correlated, respectively, with an ability of IFNgamma-activated macrophages to equivalently control strains 2308 and bacA(mut)-KL7 and the ability of bacA(mut)-KL7 organism and its LPS to induce greater amounts of pro-inflammatory cytokines than 2308. We conclude that attenuation properties of bacA mutation are dependent upon the nature of the host but more importantly that bacterial gene deletion can result in increased host pathology without an increase in bacterial load, crucial considerations for vaccine design.

Full-text

Available from: Radhika Goenka
Original article
Brucella abortus bacA mutant induces greater pro-inflammatory
cytokines than the wild-type parent strain
Michelle A. Parent
a,1,2
, Radhika Goenka
a,1
, Erin Murphy
a
, Kristen LeVier
b
,
Nuno Carreiro
a,3
, Basil Golding
c
, Gail Ferguson
b,4
, R. Martin Roop II
d
,
Graham C. Walker
b
, Cynthia L. Baldwin
a,
*
a
Department of Veterinary and Animal Sciences, University of Massachusetts, Amherst, MA 01003, USA
b
Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
c
Division of Hematology, Food and Drug Administration, Bethesda, MD 20892, USA
d
Department of Microbiology and Immunology, East Carolina University School of Medicine, Greenville, NC 27858, USA
Received 15 May 2006; accepted 12 October 2006
Available online 11 December 2006
Abstract
The inner-membrane protein BacA affects Brucella LPS structure. A bacA deletion mutant of Brucella abortus, known as KL7 (bacA
mut
-
KL7), is attenuated in BALB/c mice and protects against challenge. Thus, bacA mutation was a candidate for incorporation into live attenuated
vaccines. We assessed bacA
mut
-KL7 in 2 additional mouse strains: the more resistant C57BL/6 that produces interferon-g throughout the infec-
tion and the highly susceptible interferon-g-deficient C57BL/6 in which brucellae exhibit continual exponential growth. While it was hypoth-
esized that bacA
mut
-KL7 would exhibit even greater attenuation relative to its parent strain B. abortus 2308 in C57BL/6 mice than it did in
BALB/c mice, this was not the case. Moreover, it was more pathogenic in C57BL/6 interferon-g-deficient mice than 2308 causing abscesses
and wasting even though the splenic loads of bacA
mut
-KL7 were significantly lower. These 2 observations were correlated, respectively, with
an ability of IFNg-activated macrophages to equivalently control strains 2308 and bacA
mut
-KL7 and the ability of bacA
mut
-KL7 organism
and its LPS to induce greater amounts of pro-inflammatory cytokines than 2308. We conclude that attenuation properties of bacA mutation
are dependent upon the nature of the host but more importantly that bacterial gene deletion can result in increased host pathology without
an increase in bacterial load, crucial considerations for vaccine design.
Ó 2006 Elsevier Masson SAS. All rights reserved.
Keywords: Brucella; Brucellosis; Intracellular bacteria; bacA
1. Introduction
Brucella abortus causes serious chronic infections in live-
stock and humans [1,2]. It is a facultative intracellular gram-
negative bacterium that resides and replicates inside host
cell compartments where it resists low pH [3], reactive oxygen
intermediates and nitric oxide [4]. The unusual lipid A of
B. abortus lipopolysaccharide (LPS) is involved in maintain-
ing the integrity of the bact erial cell envelope [5] and proposed
to be important for immune evasion during the early stages of
Brucella infection [6]. In support of a role for lipid A in
immune evasion, a bacA mutant of B. abortus designated as
strain KL7 (bacA
mut
-KL7) has been shown to be attenuated
in BALB/c mice [7].
Abbreviations: CFU, colony forming units; IFNg, interferon-g; IL, inter-
leukin; LPS, lipopolysaccharide; PBS, phosphate-buffered saline; TLR, toll-
like receptor; TNFa, tumor necrosis factor-a.
* Corresponding author. Department of Veterinary and Animal Sciences,
University of Massachusetts, 161 Holdsworthway, Amherst, MA 01003,
USA. Tel.: þ1 413 545 3167; fax: þ1 413 545 6326.
E-mail address: cbaldwin@vasci.umass.edu (C.L. Baldwin).
1
These two authors contributed equally.
2
Present address: Trudeau Institute, Saranac Lake, NY, USA.
3
Present address: Colorado State University College of Veterinary Medi-
cine, CO, USA.
4
Present address: Edinburgh University, Edinburgh, Scotland, UK.
1286-4579/$ - see front matter Ó 2006 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.micinf.2006.10.008
Microbes and Infection 9 (2007) 55e62
www.elsevier.com/locate/micinf
Page 1
BacA is a cytoplasmic membrane transport protein with 7
predicted membrane-spanning domains [8] that is involved
in the generation of unusual very-long-chain fatty acids
(C 28) in lipid A biosynthesis by an unresolved mechanism
[9]. Mutation of bacA results in only half of the lipid A mol-
ecules having very-long-chain fatty acids. Such a modification
may compromise the integrity of the bacterial cell envelope [9]
or alter the ability of the bacteria to stimulate or withstand host
immune responses. As a result bacA mutation has been sug-
gested for inclusion in vaccine candidates [7].
Both BALB/c (the susceptible host model) and C57BL/6
(the resistant host model) mice depend upon IFNg to survive
infections with Brucella [10,11] probably because IFNg acti-
vates macrophages for increased killing of Brucella [12,13].
Despite the imperative role of IFNg in control of brucellosis,
BALB/c mice exhibit a hiatus of IFNg production during the
plateau phase of infection, whereas C57BL/6 mice produce
IFNg throughout the course of infection. The latter is correlated
with lower inf ection levels during the plateau phase and more
rapid clearance of the infection in C57BL/6 mice [11] . Since
bacA
mut
-KL7 was more susceptible to BALB/c macrophage
killing than 2308 [7], we hypothesized that bacA
mut
-KL7 also
would be more susceptible to killing by IFNg-activated macro-
phages than 2308 and thus would exhibit greater attenuation in
C57BL/6 mice than it did in BALB/c mice because of the con-
tinual IFNg production in the former. Furthermore, since bacA
deletion or disruption was proposed for inclusion in live vac-
cine candidates, it was important to evaluate its clearance in
the severely immune-compromised C57BL/6 IFNg-deficient
mice that die by 6 weeks following infection with the virulent
parent strain B. abortus 2308 [11] . Both were evaluated here.
2. Material and methods
2.1. Bacterial strains
Work with Brucella was approved by the Institutional Bio-
safety Committee and CDC (registration #C20041019-0289).
B. abortus 2308 was the parent strain for the bacA mutant
strain designated as KL7 (bacA
mut
-KL7) which was produced
by deleting a portion of the bacA gene as described [7]. Other
strains used as reference or controls where indicated included
B. abortus RB51, Brucella melitensis 16M, a bacA mutant of
B. melitensis 16M (designated strain KL20) and the host-factor
I deletion mutant of B. abortus known as hfq3 (generated as
described [14]).
2.2. Bacterial strain confirmation
BacA mutants are sensitive to bleomycin. To test for sensi-
tivity, B. abortus 2308 and bacA
mut
-KL7 were streaked onto
Brucella agar following the standard KirbyeBauer procedure
[15]. Filter paper discs were inoculated with 0.03 U of bleo-
mycin sulfate (Sigma, St. Louis, MO) diluted in dimethylsulf-
oxide and phosphate-buffered saline (PBS) and then dropped
onto the streaked agar plates. Control discs were inoculated
with dimethylsulfoxide and PBS. Zones of inhibition of bacte-
rial growth were measured after 72 h.
To confirm bacterial species, DNA was isolated from sam-
ples of B. abortus bacA
mut
-KL7 and 2308 and B. melitensis
16M and KL20 as described [16]. PCR amplification of the
IS711 elements [17] and Southern blotting was performed as
described [18]. For Southern blotting, DNA was digested
with EcoRI, electrophoresed, transferred to nitrocellulose
membranes and probed with a 700 bp fragment representative
of the IS711 element.
2.3. In vivo infection and post-infection analysis
BALB/c, C57BL/6 and C57BL/6 IFNg/ mice
(B6.129S7-Ifng
tm1Ts
) were purchased from Jackson Laborato-
ries (Bar Harbor, Maine) and maintained in barrier housing
in an ABSL3. Animal experiments were approved by the Uni-
versity of Massachusetts’ IACUC. Eight to 12-week-old mice
were infected intra-peritoneally with 5 10
4
bacterial colony
forming units (CFU) of B. abortus 2308, bacA
mut
-KL7 or
hfq3. Clinical scores used for characterizing the health of
mice were: 1, healthy ; 2, mildly lethargic; 3, lethargic and
emaciated; 4, lethargic, emaciated and hair loss; 5, lethargic,
wasted, hair loss and abscesses.
At various times post-infection, spleens were aseptically
removed, weighed and homogenized in PBS. Total CFU was
determined by plating serial dilutions onto Brucella agar
(BBL BD Biosciences, Bedford, MA) plates. The number of
viable splenocytes was determined by microscopy usin g try-
pan blue exclusion. Analysis of viable splenocyte populations
was conducted by immunofluorescence using monoclonal an-
tibodies against CD4 (L3T4), CD8 (Ly-2a) and Mac1 (M1/70)
(Pharmingen, San Diego, CA) and analysis by flow cytometry
using standard techniques.
2.4. Macrophages and infection
C57BL/6 peritoneal macrophages were induced by intra-
peritoneal injection of mice with 5% protease peptone. Macro-
phages were collected by lavage 5e7 days later, suspended in
medium consisting of RPM I 1640 supplemented with 10%
heat-inactivated fetal bovine serum and 2 uM
L-glutamine.
They were infected with bacA
mut
-KL7 or 2308, in the absence
or presence of exogenous IFNg (Roche, Nutley, NJ) as de-
scribed [4]. Briefly, after a 2 h infection extracellular bacteria
were removed and gentamicin at 100 mg/ml added until the
time of assessment. After various times of culture, gentamicin
was removed and macrophages lysed by addition of 0.1%
deoxycholate and CFU/culture determined by serial dilution
plating onto agar plates.
2.5. Complement sensitivity assay
B. abortus strains RB51, 2308 or bacA
mut
-KL7 organisms
were incubated with undiluted fresh or heat-inactivated
(56
C for 1 h) bovine serum for 24 h at 37
C as described
56 M.A. Parent et al. / Microbes and Infection 9 (2007) 55e62
Page 2
[19]. On termination of the assay, CFU in triplicate cultures
was determined by plating serial dilutions.
2.6. LPS isolation
LPS was purified from B. abortus strains bacA
mut
-KL7 and
2308 using the hot-phenol method (Redf earn Ph.D. thesis,
University of Wisconsin-Madison, Madison, Wis., 1960).
LPS was evaluated for carbohydrate content by Periodic
Acid Schiff (PAS) staining following electrophoresis on 10%
SDS-PAGE gels and LPS was quantified by Purpald reagent
(Sigma) as described [20]. As a comparison, enterobacteriacea
LPS isolated from Salmonella enterica serovar minnesota by
phenol water extraction (Sigma) was used.
2.7. Cytokine assays
Protease peptone-elicited peritoneal macrophages were cul-
tured with medium (described above) alone, or with medium
containing heat-killed B. abortus strains 2308 or bacA
mut
-
KL7 at a ratio of bacteria: macrophage of 3000:1, or in sepa-
rate experiments with 0.1 mg/ml S. enterica serovar minnesota
LPS (Sigma) or 100 mg/ml of 2308 or bacA
mut
-KL7 LPS
diluted in medium. The culture supernatants were harvested
after 48 h. For serum cytokine levels, 3e6 10
7
heat-killed
organisms of each bacteri al strain was injected intra-
peritoneally into C57BL/6 mice and the serum was obtained
at 2 h post-injection as described elsewhere [21]. Standard
sandwich ELISAs were established to measure interleukin-
12 (IL-12)p40 and IL-10 using capture and detection anti-
bodies and recombinant cytokines to generate standard curves
(Pharmingen) while the tumor necrosis factor- a (TNFa)
ELISA was conducted using a commercially-available kit
and following the manufacturer’s instructions (Pierce Endo-
gen, Rockford, IL). All cultures for supernatants were set-up
in triplicate and each supernatant was evaluated in replicate
wells by ELISA.
2.8. Statistical analysis
Data were subject ed to Student’s t-test using Minitab Statis-
tical software (Minitab Inc., State College, PA).
3. Results
3.1. BacA deletion mutant is not more attenuated
in resistant C57 BL/6 mice
It was hypothesized that attenuation resulting from deletion
of the bacA gene in B. abortus would be more pronounced in
C57BL/6 mice than had been shown for BALB/c mice since
C57BL/6 mice mount a strong T
H
1 immune response to infection
with B. abortus, characterized by continuous IFNg production
[22]. Interestingly, at 4 weeks post-infection,a time corresponding
to the plateau phase of infection, there was no significant differ-
ence in the number of 2308 and bacA
mut
-KL7 CFU recovered
from C57BL/6 mice (Fig. 1A). However, bacA
mut
-KL7 was sig-
nificantly attenuated in BALB/c mice during this phase
(>1og
10
2 fewer CFU than 2308) as reported previously [7]
(Fig. 1A). During the clearance phase of the infection bacA
mut
-
KL7 did show significant attenuation in C57BL/6 mice relative
to 2308 (Fig. 1B) although it was much less than that observed
in BALB/c mice (log
10
1.2 fewer bacA
mut
-KL7 CFU recovered
from C57BL/6 mice but >log
10
2 fewer bacA
mut
-KL7 CFU
recovered from BALB/c mice relative to 2308 CFU) (Fig. 1B).
A B. abortus hfq mutant was evaluated as an attenuation
control in C57BL/6 mice since it also is significantly attenuated
in BALB/c mice [14]. The hfq gene product, host-factor I, is
an RNA-binding protein necessary for the post-transcriptional
enhancement of rpoS [14] , a known sigma factor in Escherichia
coli. Hfq protein is required for transcription of genes involved
in stationary phase of growth whose products protect bacteria
from a variety of environmental stresses such as those encoun-
tered when establishing chron ic infections in mammalian
hosts. Because of the predicted effect of hfq gene deletion on
a large number of genes it was expected to be severely attenu-
ated in all mouse strains. Indeed at least log
10
3.4 fewer Hfq3
CFU than 2308 CFU were recovered from C57BL/6 mice at
both phases of the infection (Fig. 1A and B).
The weight and leukocyte composition of the spleens
of C57BL/6 mice infected with bacA were also evaluated at
6 weeks post-infection. These parameters were not signifi-
cantly different from those in mice infected with the parent
strain 2308 (Table 1). In contrast, Hf q3-infected mice had sig-
nificantly smaller spleens as well as significantly fewer total
BALB/c C57BL/6
(A) (B)
BALB/c
Detection Limi
t
*
3/5 mice
no growth
C57BL/6
*
2308
KL7
hfq3
*
0
1
2
3
4
5
6
7
CFU/Spleen (log 10)
0
1
2
3
4
5
6
7
CFU/Spleen (log 10)
*
4/5 mice
no growth
*
Plateau phase Clearance phase
Fig. 1. Infection of wild-type C57BL/6 or BALB/c mice with B. abortus strains bacA
mut
-KL7, 2308 or Hfq3. (A) Plateau phase was 4 weeks post-infection while
(B) initiation of clearance phase is 6 weeks post-infection for C57BL/6 mice and 8 weeks for BALB/c mice. Mean SEM for groups of 5 mice shown, repre-
sentative of 2 experiments. Significant decrease P 0.05 in recovery of mutant bacterial strains relative to 2308 are indicated by asterisks.
57M.A. Parent et al. / Microbes and Infection 9 (2007) 55e62
Page 3
CD4
þ
, CD8
þ
and MAC-1
þ
cells at 6 weeks post-i nfection
(Table 1) indicating less severe inflammation and/or a reduced
immune response.
To confirm that the recovered bacteria were the B. abortus
KL7 strain (where expected), PCR, Southern blotting and sen-
sitivity to bleomycin were employed with B. abortus strain
2308, B. melitensis 16M and a bacA deletion mutant of B. me-
litensis 16M cal led strain KL20 as controls in this experiment.
The results confirmed the species and gene mutation for the re-
covered organisms (data not shown). Thus, it can be concluded
from experiments reported here that the attenuation due to par-
tial deletion of the bacA gene in B. abortus was not enhanced
in the continual presence of host IFNg. Rather, bacA
mut
-KL7
was more attenua ted in BALB/c mice suggesting that bacA
mut
-
KL7 is more susceptible to the mechanisms of control that pre-
vail in BALB/c mice relative to 2308 [23].
3.2. BacA
mut
-KL7 is more pathogenic in IFNg-deficient
mice than the parent strain
There is a continual increase in CFU of B. abortus 2308 fol-
lowing infection of C57BL/6 IFNg gene-deleted mice
(IFNg/ mice) [11] and, hence, no plateau or clearance
phase. Control of bacA
mut
-KL7 was evaluated in these mice
to determine the contribution of IFNg-based immune pressure
to its attenuation. We recovered significantly more B. abortus
2308 and bacA
mut
-KL7 CFU from C57BL/6 IFNg-deficient
mice (Fig. 2) relative to that recovered from wi ld-type
C57BL/6 mice (refer to Fig. 1) at both times post-infection.
This observation is consistent with previous results for 2308
[11] and demonstrates a role for IFNg in control for both
2308 and bacA
mut
-KL7. Moreover, there were significantly
fewer bacA
mut
-KL7 CFU than 2308 CFU recovered from
C57BL/6 IFNg/ mice at both times evaluated ( Fig. 2).
While this latter result indicates that bacA
mut
-KL7 is attenu-
ated in these mice, the B. abortus hfq3 mutant was much
more severely attenuated. There was log
10
0.4e0.6 fewer
CFU of bacA
mut
-KL7 at 4 and 6 weeks post-infection than
2308 while the hfq3 mutant had a mean of log
10
4.3 fewer
CFU at 4 weeks post-infection and no hfq3 CFU were detect-
able at 6 weeks (Fig. 2).
While spleen weights of bacA
mut
-KL7 infected IFNg-
deficient mice were not significantly different from those of
2308-infected mice in either experiment, spleen weights of
hfq3-infected mice were lower than those from 2308-infected
mice with statistically significant differences at 6 weeks post-
infection (Table 1). Flow cytometric analysis of splenocytes
showed that mice infected with bacA
mut
-KL7 had significantly
fewer CD4
þ
T cells and MAC-1
þ
cells at 4 weeks post-infection
than did 2308-infected mice, while at 6 week post-infection
they had significantly fewer CD8
þ
T cells. Spleens from
hfq3-infected mice had significantly fewer total cells (data
not shown), CD4
þ
and CD8
þ
T cells and MAC-1
þ
cells at
Table 1
Splenic composition of C57BL/6 mice infected with B. abortus strains
a
Mouse
strain
B. abortus strain Week
post-infection
Spleen
weight (mg)
Number of cells/spleen (l0
7
)
b
CD4 CD8 MAC-1
Wild-type 2308 4 ND
c
5.7 (1.8) 4.4 (0.6) 6.0 (2.0)
bacA
mut
-KL7 106 (12) 5.2 (0.6) 5.6 (0.8) 4.5 (0.6)
Hfq3 98 (5) 5.2 (0.4) 5.9 (0.6) 5.7 (0.9)
2308 6 178 (38) 3.2 (0.3) 3.1 (0.4) 2.1 (0.7)
bacA
mut
-KL7 143 (8) 4.7 (0.5) 3.2 (0.3) 1.5 (0.2)
Hfq3 69 (2)* 1.7 (0.4)** 1.3 (0.3)** 0.57 (0.1)*
IFNg/ 2308 4 875 (40) 19.0 (1.7) 11.1 (0.8) 48.0 (4.4)
bacA
mut
-KL7 745 (42) 11.3 (2.1)* 6.3 (1.9) 22.0 (3.4)**
Hfq3 525 (188) 0.3 (0.1)* 0.2 (0.1)* 0.3 (0.1)*
2308 6 783 (121) 15.5 (2.7) 8.7 (0.6) 35.3 (4.8)
bacA
mut
-KL7 1018 (50) 11.5 (2.2) 3.3 (0.8)** 30.4 (0.8)
Hfq3 124 (18)* 3.1 (0.3)* 3.0 (0.3)** 2.4 (0.2)**
*Statistically significant at P 0.05; and **statistically significant at P 0.01 relative to that in the control group infected with B. abortus 2308.
a
Results presented are mean (SEM), n ¼ 5.
b
Total cells per spleen were determined by multiplying the percentage of each cell population determined by immunostaining and flow cytometry analysis by the
total number of splenocytes recovered.
c
ND, not determined due to technical failure.
0
1
2
3
4
5
6
7
8
Time post-infection
(
weeks
)
CFU/Spleen (log 10)
2308
KL7
hfq
4 6
2/5
no growth
5/5
no growth
Detection Limit 0.3 log
*
*
*
Fig. 2. Infection of C57BL/6 IFNg/ mice with B. abortus strains 2308,
bacA
mut
-KL7 or Hfq3. CFU determined at 4 and 6 weeks post-infection with
mean SEM for groups of 5 mice shown, representative of 2 experiments.
Significant decrease P 0.05 in recovery of mutant bacterial strains relative
to 2308 are indicated by asterisks.
58 M.A. Parent et al. / Microbes and Infection 9 (2007) 55e62
Page 4
both 4 and 6 weeks post-infection indicative of reduced in-
flammation (Table 1).
Despite the decreased CFU recovered from mice infected
with bacA
mut
-KL7 relative to that from mice infected with
2308, interestingly, they had striking differences in pathology.
The bacA
mut
-KL7-infected mice were characterized by being
moribund and wasted with abscesses on their tails. In contrast,
the 2308-infected IFNg-deficient mice were mildly lethargic with
matted fur. That is, on a clinical scale with 1 being normal and
5 moribund and wasted with abscesses, 40% of the bacA
mut
-
KL7-infected mice scored 5 by 3 weeks post-infection and
100% scored 5 at both 4 and 5 weeks post-infection, while
the 2308-infected mice scored 2 throughout this time. There
was no difference in survival rates as 80% of mice survived
until 6 weeks post-infection in both groups. In addition, the
spleens of the bacA
mut
-KL7-infected mice were friable at
the time of necropsy. The Hfq3-infected C57BL/6 IFNg-defi-
cient mice included as a control were all active and grooming,
scoring 1 on the clinical scale, indicating that attenuation with
this strain of Brucell a is independent of substantial immune
pressure. To conclude, in comparison to the parent strain
2308, bacA
mut
-KL7 induced greater levels of pathology
when the host was unable to control the infection, i.e. in the
IFNg-deficient mice.
3.3. Susceptibility of bacA
mut
-KL7 by IFNg-activated
macrophages
Since C57BL/6 mice depend upon continual IFNg pressure
for control of brucellosis, we hypothesized that the decreased
attenuation of bacA
mut
-KL7 relative to 2308 in this mouse
strain compared to that observed in BALB/c mice might reflect
a difference in the ability of C57BL/6 macrophages to control
bacA
mut
-KL7 relative to 2308. When tested, C57BL/6 macro-
phages exhibited a significantly greater ability to kill bacA
mut
-
KL7 at 24 and 48 h post-infection compared to 2308 (Fig. 3A).
Addition of IFNg to the macrophage cultures increased killing
of 2308 by 24 h and prevented subsequent replication of both
2308 and bacA
mut
-KL7 with no significant difference in the
recovery of these two B. abortus strains at 48 h (Fig. 3B).
Hence, we conclude that bacA
mut
-KL7 is more sensitive to
the C57BL/6 macrophage killing than 2308, consistent with
previous reports with BALB/c macrophages, but 2308 and
bacA
mut
-KL7 are equally susceptible to control by IFNg-
activated macrophages. Thus, the slightly increased sensitivity
to macrophage killing may contribute to the lower CFU of
bacA
mut
-KL7 recovered from both wild-type and IFNg-
deficient C57BL/6 mice during the clearance ph ase of the
infection. However, the ability of the IFNg-activated macro-
phages to equivalently control 2308 and bacA
mut
-KL7 could
result in a decr eased ability to measure bacA
mut
-KL7 attenua-
tion in wild-type C57BL/6 mice compared to what was previ-
ously measured in BALB/c mice.
3.4. BacA
mut
-KL7 is sensiti ve to complement-mediated
control
LPS from gram-negative bacteria have the ability to stimu-
late the alternative pathway of complement activation. More-
over, rough strains of Brucella that lack the complete LPS
O-polysaccharide side chain tend to be particularly sensitive
to complement-mediated mechanisms of control whereas
smooth strains are not [19]. Since bacA
mut
-KL7 has an altered
LPS molecule, albeit different from that of rough strains, we
wished to determine if it was also more sensitive to comple-
ment-mediated killing than the parent strain. This could con-
tribute to the reduced recovery of bacA
mut
-KL7 from
infected mice. Indeed, bacA
mut
-KL7 exhibited sensitivity to
complement killing which was comparable to the rough strain
B. abortus RB51 (Fig. 4). Thus, complement-mediated lysis of
extracellular bacteria could partially account for the lower
***
***
*
2308 KL7
RB51
0
2
4
6
8
serum
Heat-inactivated serum
Bacterial strains
Log
10
CFU recovered
Fig. 4. Susceptibility of B. abortus strains 2308, RB51 or bacA
mut
-KL7 to
complement killing. Bacteria were incubated with fresh bovine serum with
or without heat-inactivation and CFU subsequently determined. Mean SEM,
representative of 2 experiments performed, are shown with significant differ-
ences indicted as: *, P 0.05; and ***, P 0.001.
0
1
2
3
4
5
6
024
48
Time post-infection
(
hours
)
Log10 CFU recovered
IFNγ-activated
2308
KL7
(A) (B)
non-IFNγ-activated
0
1
2
3
4
5
6
02448
Time post-infection
(
hours
)
Log10 CFU recovered
**
*
Fig. 3. C57BL/6 macrophages were cultured with or without recombinant
IFNg, infected with B. abortus strains 2308 or bacA
mut
-KL7 and CFU deter-
mined at times indicated. Mean SEM, representative of 2 experiments per-
formed, is shown with significant differences indicated as: *, P 0.05; and
**, P 0.01.
59M.A. Parent et al. / Microbes and Infection 9 (2007) 55e62
Page 5
CFU of bacA
mut
-KL7 recovered compared to B. abortus 2308
in C57BL/6 wild-type, C57BL/6 IFNg-deficient and BALB/c
mice.
3.5. BacA
mut
-KL7 induces pro-inflammatory cytokines
It is known that loss of BacA protein results in alteration of
the lipid A component of the LPS molecule [9] although it is
unknown whether there are additional effects. It is also known
that (i) different LPS molecules interact differently with toll-
like receptors (TLR) [24], (ii) BALB/c and C57BL/6 mice dif-
fer in their TLR-mediated responses [25], and (iii) stimulation
of TLR by B. abortus LPS induces TNFa production
(a mediator of cachexia) and IL-12p40 [26,27]. In addition,
heat-killed B. abortus is known to induce IL-10 [27], an anti-
inflammatory cytokine that can dampen both TNFa and IL-12-
mediated effects. Thus, we evaluated whether bacA
mut
-KL7
differed from B. abortus 2308 in its ability to induce these cy-
tokines. To test this, peritoneal macrophages from C57BL/6
mice were cultured with heat-killed bacA
mut
-KL7, 2308 or
medium and culture supernatants were evaluated for TNFa,
IL-12p40 and IL-10. In vivo experiments were also conducted
by intra-peritoneal administration of heat-killed bacA
mut
-KL7
or 2308 or PBS and mouse sera evaluated for IL-12p40 and
TNFa. There was an overall tendency of bacA
mut
-KL7 organ-
isms to induce higher amounts of pro-inflammatory cytokines
both in vivo (Fig. 5A) and in vitro (Fig. 5B) compared to 2308:
IL-12 was produced at significantly higher amounts in vitro
while both IL-12 and TNFa were significantly higher in vivo
(Fig. 5). Furthermore, bacA
mut
-KL7 LPS induced significantly
higher TNFa than parental 2308 LPS (Fig. 5C). While this
may suggest a greater potential of bacA
mut
-KL7 LPS to induce
endotoxemia, bacA
mut
-KL7 LPS was much less potent than
Salmonella LPS; one thousand-fold less Salmonella LPS in-
duced more TNFa than bacA
mut
-KL7 LPS (Fig. 5C) corrobo-
rating earlier reports of low endotoxin potential of Brucella
LPS compared to that of classical enterobacterial LPS [28].
Moreover, the fact that B. abortus strains induced high
amounts of IL-10 may contribute to their reduced ability to in-
duce septic shock and although there was more IL-10 pro-
duced in response to bacA
mut
-KL7 than 2308, the difference
was not significant (Fig. 5B). Hence, the cytokine response eli-
cited by bacA
mut
-KL7 is consistent with the greater inflamma-
tion and immunopathology induced by bacA
mut
-KL7 in the
IFNg-deficient C57BL/6 mice.
4. Discussion
Previously it was shown that the B. abortus bacA deletion
mutant strain bacA
mut
-KL7 was attenuated in BALB/c mice
[7], which is the susceptible model for brucellosis [22].
Although we hypothesized that bacA
mut
-KL7 would be more
attenuated relative to 2308 in the more resistant C57BL/6
mice, this was not the case. It is a plausible explanation that
because C57BL/6 mice clear the parent strain 2308 more effi-
ciently than do BALB/c mice that attenua tion of bacA
mut
-KL7
is less obvious. That is, the IFNg-based immunity prevalent
in C57BL/6 mice may not be mor e effective at clearing
(A) (B)
*
TNFα
IL-12 IL-10
0
500
1000
1500
Medium
2308
KL7
2500
5000
Cytokines tested
Supernatant concentration
(pg/ml)
*
TNFα
IL-12
0
100
200
300
PBS
2308
KL7
600
1000
Cytokines tested
Serum concentration
(pg/ml)
*
(C)
*
Medium Salmonella 2308 KL7
0
500
1000
LPS concentration
(
u
g
/ml
)
0.1 100 100
Concentration of TNFα
(pg/ml)
Fig. 5. Cytokines produced in response to B. abortus 2308 or bacA
mut
-KL7. (A) C57BL/6 mice were inoculated with heat-killed B. abortus 2308 or bacA
mut
-KL7
or PBS and cytokine levels determined in serum. C57BL/6 peritoneal macrophages were cultured with (B) heat-killed B. abortus 2308 or bacA
mut
-KL7 organisms
or medium or with (C) LPS from Salmonella or B. abortus strain 2308 or bacA
mut
-KL7 and cytokine levels determined in supernatants. Data are mean SEM;
significant differences of KL7 compared to 2308 (P < 0.05) are indicated by asterisks.
60 M.A. Parent et al. / Microbes and Infection 9 (2007) 55e62
Page 6
bacA
mut
-KL7 than 2308 while the alternative immune control
mechanisms that operate in BALB/c mice during the plateau
phase of infection are (The complete mechanisms operating
in BALB/c mice are unclear but include TNFa [23]). In sup-
port of this, while bacA
mut
-KL7 was more vulnerable to kill-
ing by C57BL/6 macrophages than was 2308 the two strains
showed similar sensitivity to control by IFNg-activated macro-
phages. Thus, any differences seen with non-IFNg-activated
macrophages would be expected to be minimized in C57BL/6
mice since they continually produce IFNg during the in-
fection. Presumably this occurs in vivo since we found
bacA
mut
-KL7 is indeed vulnerable to IFNg-based mechanisms
of control shown by the higher bacA
mut
-KL7 CFU recovered
from the C57BL/6 IFNg-deficient mice than from C57BL/6
wild-type mice. In contrast, sensitivity to complement-based
killing, reported here, would be expected to contribute to de-
creased bacA
mut
-KL7 recovery in all stains of mice.
The surprising result was the greater virulence of bacA
mut
-
KL7 in immune-deficient mice. Gene deletion resulting in
increased virulence is not without precedence as it has been
reported for Shigella, E. coli and Brucella that have lost their
ability to acquire iron [29]. Here the increased inflammation in
the IFNg-deficient mice infected with bacA
mut
-KL7 is consis-
tent with the increase in pro-inflammatory cytokines induced
by bacA
mut
-KL7 and its LPS. While incr eased ability to stim-
ulate TNFa may be particularly germane to its attenuation in
BALB/c mice whose immunity to brucellosis is known to de-
pend upon TNFa as noted above [23]. The increased produc-
tion of pro-inflammatory ctyokines is likely to increase
pathogenesis in IFNg-deficient mice since the increase in
IL-12 would not result in an increase in the protective IFNg
in those mice.
The increase in pro-inflammatory cytokines also favors the
hypothesis that reduced very-long-chain fatty acid modifica-
tion of the lipid A in bacA
mut
-KL7 engenders the LPS with
greater capacity to interact with host TLR [9], albeit purified
bacA
mut
-KL7 LPS was still less potent than Salmonella LPS.
In addition, the altered LPS of bacA
mut
-KL7 may enable it
to interact with additional types of TLR [24] although both
are speculative. Finally, it cannot be ruled out that unidentified
alterations in addition to the lipid A effect could result from
the bacA gene deletion and contribute to the increased patho-
logy induced by bacA
mut
-KL7. These possibilities provide av-
enues for future investigations.
In conclusion, these data show that the degree of attenua-
tion of the bacA
mut
-KL7 is depend ent upon the host and the
nature of its immune response. Loss of bacA resulted in less
attenuation relative to 2308 when the mutant was evaluated
in C57BL/6 mice compared to that in BALB/c mice. More-
over, bacA
mut
-KL7 caused unique pathology in immune-
deficient mice characterized by abscesses and wasting. These
observations illustrate important considerations for designing
live vaccines and are in keeping with the observation that
while the live vaccine strain B. abortus S19 is attenuated in
cattle, it is not attenuated in humans. They also indicate that
a bacA mutation in a potential vaccine strain may be particu-
larly inappropriate for hosts with immune deficiencies.
Acknowledgements
This work was supported by USDA-CSREES competitive
grants program #98-02620.
References
[1] E. Moreno, E. Stackebrandt, M. Dorsch, J. Wolters, M. Busch, H. Mayer,
Brucella abortus 16S rRNA and lipid A reveal a phylogenetic relation-
ship with members of the alpha-2 subdivision of the class proteobacteria,
J. Bacteriol. 172 (1990) 3569e3576.
[2] C.L. Baldwin, R. Goenka, Host immune responses to the intracellular
bacteria Brucella: Does the bacteria instruct the host to facilitate chronic
infection? Crit. Rev. Immunol. 26 (2006) 407e442.
[3] M.G. Rittig, M.T. Alvarez-Martinez, F. Porte, J.P. Liautard, B. Rouot,
Intracellular survival of Brucella spp. in human monocytes involves con-
ventional uptake but special phagosomes, Infect. Immun. 69 (2001)
3995e4006.
[4] X. Jiang, B. Leonard, R. Benson, C.L. Baldwin, Macrophage control of
Brucella abortus: role of reactive oxygen intermediates and nitric oxide,
Cell Immunol. 151 (1993) 309e319.
[5] R.M. Roop, G.T. Robertson, G.P. Ferguson, L.E. Milford, M.E. Winkler,
G.C. Walker, Seeking a niche: putative contributions of the hfq and bacA
gene products to the successful adaptation of the brucellae to their intra-
cellular home, Vet. Microbiol. 90 (2002) 349e363.
[6] J. Goldstein, T. Hoffman, C. Frasch, E.F. Lizzio, P.R. Beining,
D. Hochstein, Y.L. Lee, R.D. Angus, B. Golding, Lipopolysaccharide
(LPS) from Brucella abortus is less toxic than that from Escherichia
coli, suggesting the possible use of B. abortus or LPS from B. abortus
as a carrier in vaccines, Infect. Immun. 60 (1992) 1385e1389.
[7] K. LeVier, R.W. Phillips, V.K. Grippe, R.M. Roop, G.C. Walker, Similar
requirements of a plant symbiont and a mammalian pathogen for pro-
longed intracellular survival, Science 287 (2000) 2492e2493.
[8] J. Glazebrook, A. Ichige, G.C. Walker, A Rhizobium meliloti homolog of
the Escherichia coli peptide-antibiotic transport protein SbmA is essen-
tial for bacteroid development, Genes Dev. 7 (1993) 1485e1497.
[9] G.P. Ferguson, A. Datta, J. Baumgartner, R.M. Roop, R.W. Carlson,
G.C. Walker, Similarity to peroxisomal-membrane protein family reveals
that Sinorhizobium and Brucella BacA affect lipid-A fatty acids, Proc.
Natl. Acad. Sci. U.S.A. 101 (2004) 5012e5017.
[10] D.M. Fernandes, X. Jiang, J.H. Jung, C.L. Baldwin, Comparison of T cell
cytokines in resistant and susceptible mice infected with virulent
Brucella abortus strain 2308, FEMS Immunol. Med. Microbiol. 16
(1996) 193 e203.
[11] E.A. Murphy, J. Sathiyaseelan, M.A. Parent, B. Zou, C.L. Baldwin, Inter-
feron-gamma is crucial for surviving a Brucella abortus infection in both
resistant C57BL/6 and susceptible BALB/c mice, Immunology 103
(2001) 511 e518.
[12] X. Jiang, C.L. Baldwin, Effects of cytokines on intracellular growth of
Brucella abortus, Infect. Immun. 61 (1993) 124e134.
[13] C.L. Baldwin, X. Jiang, D.M. Fernandes, Macrophage control of
Brucella abortus: influence of cytokines and iron, Trends Microbiol. 1
(1993) 99e104.
[14] G.T. Robertson, R.M. Roop Jr., The Brucella abortus host factor I (HF-I)
protein contributes to stress resistance during stationary phase and is
a major determinant of virulence in mice, Mol. Microbiol. 34 (1999)
690e700.
[15] E.W. Koneman, S.D. Allen, W.M. Janda, P.C. Schreckenberger,
W.C. Winn Jr. (Eds.), Color Atlas and Textbook of Diagnostic Microbi-
ology, Lippincott, Philadelphia, 1997.
[16] D.G. Pitcher, N.A. Saunders, R.J. Owen, Rapid extraction of bacterial
genomic DNA with guanidium thiocyanate, Lett. Appl. Microbiol. 8
(1989) 151 e156.
[17] B.J. Bricker, S.M. Halling, Differentiation of Brucella abortus bv. 1, 2,
and 4, Brucella melitensis, Brucella ovis, and Brucella suis bv. 1 by
PCR, J. Clin. Microbiol. 32 (1994) 2660e2666.
61M.A. Parent et al. / Microbes and Infection 9 (2007) 55e62
Page 7
[18] F. Ausubel, R. Brent, R. Kingston, D. Moore, J. Seidman, J. Smith,
K. Struhl (Eds.), Short Protocols in Molecular Biology, John Wiley &
Sons, Inc, New York, 1999.
[19] L.B. Corbeil, K. Blau, T.J. Inzana, K.H. Nielsen, R.H. Jacobson,
R.R. Corbeil, A.J. Winter, Killing of Brucella abortus by bovine serum,
Infect. Immun. 56 (1988) 3251e3261.
[20] C.H. Lee, C.M. Tsai, Quantification of bacterial lipopolysaccharides
by the purpald assay: measuring formaldehyde generated from 2-
keto-3-deoxyoctonate and heptose at the inner core by periodate
oxidation, Anal. Biochem. 267 (1999) 161e168.
[21] L.Y. Huang, J. Aliberti, C.A. Leifer, D.M. Segal, A. Sher,
D.T. Golenbock, B. Golding, Heat-killed Brucella abortus induces
TNF and IL-12p40 by distinct MyD88-dependent pathways: TNF, unlike
IL-12p40 secretion, is Toll-like receptor 2 dependent, J. Immunol. 171
(2003) 1441e1446.
[22] J.A. Montaraz, A.J. Winter, Comparison of living and nonliving vaccines
for Brucella abortus in BALB/c mice, Infect. Immun. 53 (1986) 245e251.
[23] E.A. Murphy, M. Parent, J. Sathiyaseelan, X. Jiang, C.L. Baldwin,
Immune control of Brucella abortus 2308 infections in BALB/c mice,
FEMS Immunol, Med. Microbiol. 32 (2001) 85e88.
[24] M.G. Netea, M. van Deuren, B.J. Kullberg, J.M. Cavaillon, J.W. Van der
Meer, Does the shape of lipid A determine the interaction of LPS with
Toll-like receptors? Trends Immunol. 23 (2002) 135e139.
[25] T. Liu, T. Matsuguchi, N. Tsuboi, T. Yajima, Y. Yoshikai, Differences in
expression of toll-like receptors and their reactivities in dendritic cells in
BALB/c and C57BL/6 mice, Infect. Immun. 70 (2002) 6638e6645.
[26] M.A. Campos, G.M. Rosinha, I.C. Almeida, X.S. Salgueiro, B.W. Jarvis,
G.A. Splitter, N. Qureshi, O. Bruna-Romero, R.T. Gazzinelli,
S.C. Oliveira, Role of Toll-like receptor 4 in induction of cell-mediated
immunity and resistance to Brucella abortus infection in mice, Infect.
Immun. 72 (2004) 176e186.
[27] G.H. Giambartolomei, A. Zwerdling, J. Cassataro, L. Bruno,
C.A. Fossati, M.T. Philipp, Lipoproteins, not lipopolysaccharide, are
the key mediators of the proinflammatory response elicited by heat-killed
Brucella abortus, J. Immunol. 173 (2004) 4635e4642.
[28] D.S. Weiss, K. Takeda, S. Akira, A. Zychlinsky, E. Moreno, MyD88,
but not toll-like receptors 4 and 2, is required for efficient clearance of
Brucella abortus, Infect. Immun. 73 (2005) 5137e5143.
[29] K.G. Wooldridge, P.H. Williams, Iron uptake mechanisms of pathogenic
bacteria, FEMS Microbiol. Rev. 12 (1993) 325e348.
62 M.A. Parent et al. / Microbes and Infection 9 (2007) 55e62
Page 8
  • Source
    • "This decrease seems to be necessary for intracellular adaptation of lon mutant from media culture condition. Recently, other groups have reported similar findings in virulent Brucella strain [22,23]. The recovery mechanisms and metabolic pathway proteins to establish intracellular Brucella adaptation were elucidated by a comprehensive analysis of its proteomes. "
    [Show abstract] [Hide abstract] ABSTRACT: The objective of this study was to isolate a Brucella lon mutant and to analyze the cytokine response of B. lon mutant during macrophage infection. A wild-type Brucella abortus strain was mutagenized by Tn5 transposition. From the mouse macrophage J774.A1 cells, total RNA was isolated at 0 hours, 6 hours, 12 hours, and 24 hours after infection with Brucella. Using mouse cytokine microarrays, we measured transcriptional levels of the cytokine response, and validated our results with a reverse transcriptase-polymerase chain reaction (RT-PCR) assay to confirm the induction of cytokine messenger RNA (mRNA). In host J774.A1 macrophages, mRNA levels of T helper 1 (Th1)-type cytokines, including tumor necrosis factor-alpha (TNF-α), interferon-gamma (IFN-γ), interleukin-2 (IL-2), and IL-3, were significantly higher in the lon mutant compared to wild-type Brucella and the negative control. TNF-α levels in cell culture media were induced as high as 2 μg/mL after infection with the lon mutant, a greater than sixfold change. In order to understand the role of the lon protein in virulence, we identified and characterized a novel B. lon mutant. We compared the immune response it generates to the wild-type Brucella response in a mouse macrophage cell line. We demonstrated that the B. lon mutants induce TNF-α expression from the host J774.A1 macrophage.
    Full-text · Article · Dec 2013
  • Source
    • "The strategy of B. abortus is to evade the innate immune system and persist in the host long enough to be transmitted. The bacterium contains an unusual lipid A composing the LPS molecule, which is important for evading the host immune system during the early stages of infection (Parent et al., 2007). In addition, when entering the host intracellular space, B. abortus displays several strategies to avoid the cellular killing mechanism. "
    [Show abstract] [Hide abstract] ABSTRACT: Innate immunity serves as the first line of defense against infectious agents such as intracellular bacteria. The innate immune platform includes Toll-like receptors (TLRs), retinoid acid-inducible gene-I-like receptors and other cytosolic nucleic acid sensors, nucleotide-binding and oligomerization domain-like receptors, adaptors, kinases and other signaling molecules that are required to elicit effective responses against different pathogens. Our research group has been using the Gram-negative bacteria Brucella abortus as a model of pathogen. We have demonstrated that B. abortus triggers MAPK and NF-κB signaling pathways in macrophages in a MyD88 and IRAK-4-dependent manner. Furthermore, we claimed that so far TLR9 is the most important single TLR during Brucella infection. The identification of host receptors that recognize pathogen-derived nucleic acids has revealed an essential role for nucleic acid sensing in the triggering of immunity to intracellular pathogens. Besides TLRs, herein we describe recent advances in NOD1, NOD2, and type I IFN receptors in innate immune pathways during B. abortus infection.
    Preview · Article · Oct 2012 · Frontiers in Cellular and Infection Microbiology
    • "In contrast to the lipid A of other Gram-negative bacteria, that of Brucella is not efficiently recognized by TLR4 (Lapaque et al., 2006). Infection with a Brucella bacA mutant that is deficient in the very long chain fatty acid content of lipid A results in higher inflammation and the mutant is attenuated in mice (Ferguson et al., 2004; Parent et al., 2007). Likewise, the increased amounts of underacylated lipid A species in a Brucella bvrRS mutant (Manterola et al., 2005 ) may partially account for their attenuation in the mouse model of infection (Sola-Landa et al., 1998). "
    [Show abstract] [Hide abstract] ABSTRACT: Bacteria of the genus Brucella are Gram-negative pathogens of several animal species that cause a zoonotic disease in humans known as brucellosis or Malta fever. Within their hosts, brucellae reside within different cell types where they establish a replicative niche and remain protected from the immune response. The aim of this article is to discuss recent advances in the field in the specific context of the Brucella intracellular 'lifestyle'. We initially discuss the different host cell targets and their relevance during infection. As it represents the key to intracellular replication, the focus is then set on the maturation of the Brucella phagosome, with particular emphasis on the Brucella factors that are directly implicated in intracellular trafficking and modulation of host cell signalling pathways. Recent data on the role of the type IV secretion system are discussed, novel effector molecules identified and how some of them impact on trafficking events. Current knowledge on Brucella gene regulation and control of host cell death are summarized, as they directly affect intracellular persistence. Understanding how Brucella molecules interplay with their host cell targets to modulate cellular functions and establish the intracellular niche will help unravel how this pathogen causes disease.
    No preview · Article · Feb 2012 · FEMS microbiology reviews
Show more