N-Formyl-Perosamine Surface Homopolysaccharides Hinder the ...
Transcript of N-Formyl-Perosamine Surface Homopolysaccharides Hinder the ...
1
N-Formyl-Perosamine Surface Homopolysaccharides 1
Hinder the Recognition of Brucella abortus by Mouse 2
Neutrophils 3
4
Ricardo Mora-Cartína,b, Carlos Chacón-Díazc, Cristina Gutiérrez-5
Jiméneza, Stephany Gurdián-Murilloa, Bruno Lomonteb, Esteban Chaves-6
Olartea,c, Elías Barquero-Calvoa,c*, Edgardo Morenoa,b* 7
8
Programa de Investigación en Enfermedades Tropicales, Escuela de Medicina Veterinaria, 9
Universidad Nacional, Heredia, Costa Ricaa,; Instituto Clodomiro Picado, Facultad de 10
Microbiología, Universidad de Costa Rica, San José, Costa Rica b; Centro de Investigación 11
en Enfermedades Tropicales, Universidad de Costa Rica, San José, Costa Rica c 12
13
Address correspondence to: Edgardo Moreno, [email protected] or Elías 14
Barquero-Calvo, [email protected] 15
16
Key words: Neutrophils, Brucella, Brucella abortus, brucellosis, mouse, complement, 17
lipopolysaccharide, perosamine 18
19
IAI Accepted Manuscript Posted Online 21 March 2016Infect. Immun. doi:10.1128/IAI.00137-16Copyright © 2016, American Society for Microbiology. All Rights Reserved.
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
2
Brucella abortus is an intracellular pathogen of monocytes, macrophages, dendritic cells 20
and placental trophoblasts. This bacterium causes a chronic disease in bovines and in 21
humans. In these hosts, the bacterium also invades neutrophils; however, it fails to replicate 22
and just resists the killing action of these leukocytes without inducing significant activation 23
or neutrophilia. Moreover, B. abortus causes the premature cell death of human neutrophils. 24
In the murine model, the bacterium is found within macrophages and dendritic cells at early 25
times of infection, but seldom in neutrophils. Following this, we explored the interaction of 26
mouse neutrophils with B. abortus. In contrast to human, dog and bovine neutrophils, naïve 27
mouse neutrophils fail to recognize smooth B. abortus at early stages of infection. Murine 28
normal serum components do not opsonize smooth Brucella strains and neutrophil 29
phagocytosis is only achieved after the appearance of antibodies. Alternatively, mouse 30
normal serum is capable to opsonize rough Brucella mutants. Despite of this, neutrophils 31
still fail to kill Brucella, and the bacterium induces cell death of these murine leukocytes. In 32
addition, mouse serum does not opsonize Yersinia enterocolitica O:9, a bacterium 33
displaying the same surface polysaccharide antigen as smooth B. abortus. Therefore, the 34
lack of murine serum opsonization and absence of murine neutrophil recognition is specific 35
and the molecules responsible for the Brucella camouflage are N-formyl-perosamine 36
surface homopolysaccharides. Although the mouse is a valuable model for understanding 37
the immunobiology of brucellosis, direct extrapolation from one animal system to another 38
has to be taken thoughtfully. 39
40
41
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
3
Polymorphonuclear neutrophil leukocytes (PMNs) are the first line of defense of the innate 42
immune system. These cells detect microbial structures through various receptors and the 43
recognition of these structures influences their activation and fate, which are essential to 44
promote inflammatory responses and host defense mechanisms. 45
Brucellosis is a chronic disease of domestic and wildlife mammals and a worldwide human 46
zoonosis caused by Brucella species (1). Members of this genus are intracellular pathogens 47
that invade monocytes (Mo), macrophages (Mφ) and dendritic cells (DCs), as well as 48
placental trophoblasts (1). Although in the natural host and in humans, Brucella also invades 49
PMNs (2-4), the bacterium fails to replicate in these cells and just resists their killing action 50
without inducing significant activation (3-6). Indeed, Brucella infected PMNs do not 51
degranulate, induce low level of reactive oxygen species (ROS) and cytokines, and the 52
infection courses without significant neutrophilia (5, 7, 8). Likewise, the PMNs infiltrate in 53
the cervical lymph nodes after oral infection is very low; even in well-developed granulomas 54
after 15 days of infection (9). Moreover, after five days of infection, the bacterium is found 55
within Mφ and DCs of mice, but seldom inside PMNs in the target organs (10). In addition, 56
the absence of PMNs during brucellosis promotes the activation of the Th1 adaptive immune 57
response (11). More significantly, Brucella is capable to induce the premature cell death of 58
human PMNs by means of its lipopolysaccharide (Br-LPS) through a mechanism that 59
involves CD14 and mild NADPH oxidase activation (6). This has led to the proposal that 60
Brucella infected PMNs function as “Trojan horses” after the non-phlogistic phagocytosis 61
by Mφ and DCs. This would favor the bacterial spreading to different organs, fostering the 62
chronicity of the disease (6). Regarding this, the mouse has been the preferred animal model 63
in brucellosis research to test and evaluate different hypotheses (12, 13). 64
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
4
Here, we have explored the interaction of naïve mouse PMNs with B. abortus and found 65
that these cells fail to recognize this bacterium in the absence of antibodies. This is 66
significant, since the outcome of brucellosis in a given animal species may be determined 67
during the initial stages of the infection that influence the downstream events of the 68
immune response. 69
70
MATERIALS AND METHODS 71
72
Ethics. Experimentation in mice was conducted with the consent and guidelines established 73
by the ‘‘Comité Institucional para el Cuido y Uso de los Animales de la Universidad de 74
Costa Rica” (CICUA-47-12), and with the corresponding law ‘‘Ley de Bienestar de los 75
Animales’’ of Costa Rica (Law 7451 on Animal Welfare). Mice were accommodated in the 76
animal building at the Veterinary Medicine School of the National University, Costa Rica. 77
All animals were kept in cages with food and water ad libitum under biosafety containment 78
conditions, previous to and during the experiment. Blood from roaming dogs kept at the 79
shelter of the Hospital of the Veterinary Medicine School of the National University, Costa 80
Rica, was taken for routine diagnostic purposes, following the respective consent and 81
approval (SIA 0434-14) and the guidelines established by ‘Ley de Bienestar de los 82
Animales’’ of Costa Rica (Law 7451 on Animal Welfare). 83
Human blood samples were collected from volunteer donors at Tropical Disease Research 84
Program (PIET), of the National University, Costa Rica, following the corresponding 85
approval (SIA 0248-13). Accordingly, volunteers were carefully informed regarding the 86
study and provided written consents. Samples were taken following the procedures dictated 87
by the Costa Rican National Health system (Ley 9234, Costa Rica, La Gaceta 79, 2014) and 88
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
5
the World Medical Association Declaration of Helsinki: Ethical Principles for Medical 89
Research Involving Human Subjects, General Assembly, Seoul, October 2008, regarding the 90
use of blood samples. 91
92
Experimental animals. Wild type Mus musculus were captured from the grounds of the 93
university campus of the National University, Costa Rica. C57BL/6, Balb/c and CD-1 mice 94
(18-21g), were provided by the following animal facilities: “Instituto Clodomiro Picado”, 95
University of Costa Rica; School of Veterinary Medicine, National University, Costa Rica 96
and; “Laboratorio de Ensayos Biológicos” from the University of Costa Rica. 97
98
Bacterial strains and Br-LPS preparations. Virulent B. abortus 2308, B. abortus 2308-99
∆wadC, B. abortus 2308-∆per (14), B. abortus 2308-GFP (15), transgenic B. abortus 2308-100
RFP with an integrated chromosomal gene coding for the red fluorescent protein from 101
Discosoma coral (provided by Dr. Jean-Jacques Letesson; Unité de Recherche en Biologie 102
Moléculaire, Facultés Universitaires Notre-Dame de la Paix, Namur, Belgium), Yersinia 103
enterocolitica O.9 (16) Staphylococcus aureus (ATCC 25923) and Escherichia coli (ATCC 104
25922) were grown in tryptic soy broth as previously described (15). Purified Br-LPS 105
suspensions were prepared from B. abortus 2308, as reported elsewhere (17). 106
107
Immunization and immune serum production against B. abortus. C57BL/6 and CD-1 108
mice were intraperitoneally infected with 0.1 mL of PBS containing 106 CFUs of virulent B. 109
abortus 2308. Mice were bled at different times, and antibody titration carried out by 110
microagglutination in 96-well round bottom plastic plates. Briefly, Rose Bengal antigen was 111
diluted 1/20 in PBS and used as a bacterial suspension for agglutination. Volumes of 50 μL 112
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
6
of serum dilution were added to 50 μL of antigen suspension. Samples were incubated at 113
4°C for 24h and bacterial agglutination recorded in the bottom of the plate. Agglutination 114
titers beyond 1/50 were considered positive. 115
For immune serum production, mice were infected as described above and after thirty days 116
of infection mice were bled and serum was separated by centrifugation and filtrated through 117
0.2 μm membrane (Millipore). Serum was then stored at -20oC in aliquots. IgGs were 118
purified from immune mouse serum as reported elsewhere (18). Western blotting analysis of 119
mouse immune serum revealed that most of the antibody recognition was directed against 120
Br-LPS (16). For ex-vivo opsonization experiments sub-agglutinating doses of antibodies 121
were added to each well. In all experiments non-immune mouse immunoglobulins were used 122
as controls and administered to the same concentration as the specific antibodies. 123
124
Bone marrow derived PMNs. Murine bone marrow cells were isolated essentially as 125
described by Boxio et al., (19). Briefly, bone marrow was collected from the femurs and 126
tibiae of Balb/c mice and suspended in Hank's Balanced Salt Solution (HBSS, no calcium, 127
no magnesium) containing 2mM EDTA and 2% inactivated fetal calf serum. After one wash 128
with 2 mL HBSS cell concentration was determined with a Neubauer chamber and PMN 129
percentage calculated by Giemsa coloration after cytospin centrifugation (Shandon Cytospin 130
2) or by flow cytometry using Guava easyCyte (Millipore). Data was analyzed with FlowJo 131
software version 10.0.7 (Tree Star, Inc.) as described (6). For some experiments, direct 132
observation of spleen cells from B. abortus 2308-GFP infected mice were performed as 133
described elsewhere (20). 134
135
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
7
PMN phagocytosis assay. Aliquots of 350 μL of human, canine or murine fresh heparinized 136
blood were incubated with bacteria or latex beads ∼2μm (Sigma-Aldrich) at the indicated 137
MOI at 37°C during one hour under mild agitation. Alternatively, bone marrow mouse 138
PMNs suspended in HBSS or mouse serum were incubated with bacteria at the indicated 139
MOI at 37°C during one hour under mild agitation. Blood smears in three glass slides were 140
fixed with methanol, centrifuged in cytospin, mounted with ProLong Gold Antifade Reagent 141
with DAPI (Thermo Fisher Scientific) and observed under the fluorescence microscope. 142
Before cell staining on glass slides, bone marrow cell suspensions were fixed with BD 143
FACS Lysing Solution or paraformaldehyde 3.5%. At least fifty PMNs were counted per 144
slide and the number of intracellular fluorescent bacteria/PMN determined. 145
146
Ex vivo bactericidal activity of serum and PMNs. Bacteria 106 colony forming units 147
(CFU) were incubated with 350 μL of normal, heat (56 °C for 30 minutes) and yeast 148
consumed and inactivated (37 °C for 1 hour) serum, or immune mouse serum at 37 ºC at 149
different times. After incubation, aliquots were dispersed on trypticase soy agar plates, 150
incubated at 37ºC for 48-72 hours and CFU determined. Aliquots of 350 μL of fresh mouse 151
heparinized blood, or bone marrow PMNs (suspended in HBSS or mouse serum) were 152
mixed at MOI of 2 or 5 bacteria/PMN under mild agitation at 37 ºC for 90 minutes. In some 153
cases the blood was supplemented with 0.25% of anti-Brucella murine immune serum. After 154
incubation, cells were lysed with 0.2% Triton X-100, 0.5 mg/mL DNAse (Sigma) and 155
aliquots dispensed on trypticase soy agar plates, incubated at 37ºC for 48-72 hours and CFU 156
determined. 157
158
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
8
PMNs cell death assays. Aliquots of 350 μL of fresh human or mouse heparinized blood 159
were mixed with different concentrations of B. abortus or Br-LPS under mild agitation at 160
37ºC for 2 hours. After incubation, red blood cells were lysed by mixing 100 µL of 161
heparinized blood with 1500 µL of red blood cell lysis buffer (NH4Cl 8.02 g, NaHCO3 0.84 162
g and EDTA 0.37 g/L, pH 7.2) for five minutes. Then the remaining leukocytes were 163
washed with ice cold PBS to remove cell debris and resuspended in 100 µL of Annexin V 164
Binding Buffer (Life technologies). Volumes of 5 µL of Annexin V and AquaDead 165
(Invitrogen) diluted 1/20 in PBS were added and incubated for 30 minutes on ice in the dark. 166
Cells were washed once with ice cold PBS, resuspended in 500 µL of BD FACS Lysing 167
solution. Samples were then subjected within one hour to flow cytometry analysis using 168
Guava easyCyte (Millipore) and data were analyzed using FlowJo software version 10.0.7 169
(Tree Star, Inc.) as described (6). 170
171
Adsorption of serum components by B. abortus and protein identification. 172
For serum adsorption 5 ×1010 CFU of the corresponding B. abortus strain were incubated 173
with 3 mL of murine or human serum at 37°C for 45 minutes under mild agitation. Control 174
bacteria were incubated only with PBS pH 7.2. Bacteria were washed 4 times with PBS pH 175
7.2. All bacterial preparations were treated with 0.1M glycine-HCl (pH 2.7) to remove 176
adsorbed proteins or with PBS for control purposes. The eluted supernatants were 177
neutralized with 1M Tris-HCl (pH 9), and precipitated with methanol-chloroform. Samples 178
were subjected to a 10% SDS-PAGE under reducing conditions. The Coomassie blue-179
stained protein bands were excised and subjected to reduction, alkylation, and in-gel tryptic 180
digestion followed by MALDI-TOF-TOF analysis on a Proteomics Analyzer 4800 Plus 181
mass spectrometer (Applied Biosystems), as described elsewhere (21). Resulting 182
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
9
fragmentation spectra were searched against the corresponding mouse or human UniProt 183
databases (www.uniprot.org) using ProteinPilot® v.4.0 and the Paragon® algorithm 184
(ABsciex) for protein identification at ≥95% confidence. 185
186
Complement and fibronectin detection. For Western blotting analysis, samples were 187
transferred to a PVDF membrane after SDS-PAGE. The membranes were blocked and 188
incubated for the detection of murine Complement C3 with 1:500 diluted Goat anti-mouse 189
C3 antibody (Thermo Scientific) and further with 1:1000 protein G-peroxidase (Life 190
Technologies). For the detection of murine fibronectin membranes were incubated with 191
1:500 diluted goat IgG anti-fibronectin antibodies (Sigma) and further with 1:1000 protein 192
G-peroxidase (Life Technologies). Proteins were detected with enhanced 193
chemiluminescence (Roche). 194
195
Statistics. ANOVA or Student’s t test were used to determine statistical significance in the 196
different assays (JASP software 2015, https://jasp-stats.org/). Data were processed in 197
Microsoft Office Excel 2013 198
199
RESULTS 200
201
B. abortus is not recognized by naïve murine PMNs. 202
We have previously shown that B. abortus is phagocytized by human PMNs and induces the 203
premature death of these leukocytes through the action of its Br-LPS (6). This has led to the 204
hypothesis that infected PMNs function as “Trojan horses” for spreading the Brucella 205
infection in different organs (6). Following this, we asked the question whether B. abortus 206
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
10
and its Br-LPS could induce the same effect in murine PMNs. As presented in Fig. 1, while 207
human PMNs died after contact with B. abortus or its Br-LPS, mouse PMNs did not show 208
any signs of cell death after exposure to these components. 209
In order to understand this phenomenon, we then explored the early interaction of murine 210
PMNs with B. abortus, and compared with that of PMNs of other animals. It has been 211
shown that naïve mouse PMNs readily ingest S. aureus and E. coli in the presence or 212
absence of complement opsonization (22, 23). In agreement with this, we also observed 213
phagocytosis of S. aureus and E. coli by PMNs of outbred CD-1 strain and wild Mus 214
musculus (Fig 2A). Likewise, human and dog PMNs readily ingested these two bacterial 215
species. In contrast to PMNs of humans and dogs, murine PMNs, failed to ingest B. abortus 216
(Fig. 2A). Phagocytosis of Brucella by murine PMNs was seldom detected even at high 217
MOI (∼100). Moreover, at these concentrations, some bacteria may lay on the cell surface 218
without internalization (Fig. 2B). The absence of Brucella recognition was specific, since 219
murine PMNs were not only capable to ingest other bacteria, but large numbers of latex 220
beads (Fig. 2B). Similar observations were obtained with PMNs from inbred C57BL/6 (Fig. 221
2B) and Balb/c mice. Consistent with previous results (5), B. abortus was isolated from the 222
blood of infected mice as early as one hour after infection, and bacteremia persisted for at 223
least two days. In spite of this, we did not detect Brucella-infected PMNs in blood or in 224
target organs during the first four days of infection, corroborating previous results (10). 225
Altogether these results demonstrate that the interaction of murine PMNs with B. abortus 226
significantly departs from that of PMNs from humans, guinea pigs, cows, goats and dogs, 227
which are capable of phagocytizing Brucella in the absence or presence of complement (3, 228
4, 24-26). 229
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
11
230
Mouse PMNs ingest B. abortus after the development of adaptive immunity. 231
Antibodies are well known opsonizing elements. We then explored the ability of murine 232
PMNs to phagocytize B. abortus during development of adaptive immune response. At the 233
onset of infection (first two days) Brucella was not internalized by murine PMNs (Fig. 3A); 234
however, at later times the bacterium was readily phagocytized by these leukocytes (Fig. 235
3A). This internalization correlated with the quick rising of murine antibodies against Br-236
LPS (the main Brucella antigen) (27). Both, immune mouse serum and specific IgG anti-237
Brucella promoted the phagocytosis of B. abortus by murine PMNs (Fig. 3B and 3C). As 238
indicated before (5), under the microscope, these infected leukocytes did not show obvious 239
signs of alterations at early times of infection (<1 hour). These results indicate that 240
phagocytosis of B. abortus by murine PMNs is efficiently mediated by opsonizing 241
antibodies through Fc receptors at the surface of these leukocytes. 242
243
Surface N-formyl-perosamine homopolysaccharides are responsible for Brucella 244
camouflage. 245
Since innate mouse opsonins failed to promote the phagocytosis of smooth B. abortus by 246
murine PMNs, then we explored the presence of blocking components on the bacterial 247
surface. First, we tested the ability of murine PMNs to phagocytize rough B. abortus 2308-248
∆per mutant lacking surface N-formyl-perosamine sugars (Br-LPS O-chain and native 249
hapten polysaccharide, known as NH) and smooth B. abortus 2308-∆wadC mutant 250
displaying a defect in core oligosaccharide (Fig. 4). While the core ∆wadC mutant was not 251
recognized by murine PMNs and behaved as the parental strain, the rough ∆per mutant was 252
readily phagocytized in the presence of normal but not inactivated mouse serum (Fig. 5A). 253
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
12
High numbers of intracellular rough B. abortus ∆per did not cause obvious alterations in 254
mouse PMNs at early times (<1 hour) of infection (Fig. 5A, insert), paralleling the results 255
obtained with antibody opsonized smooth brucellae (Fig. 3C). 256
To confirm the role N-formyl-perosamine sugars in the Brucella camouflage, then we tested 257
the phagocytosis of Yersinia enterocolitica O:9. The O-chain and NH surface molecules of 258
this bacterium are identical to those of B. abortus (16, 28). Y. enterocolitica O:9 was not 259
phagocytized by mouse PMNs in the presence of normal mouse serum (Fig. 5B). However, 260
this bacterium was readily internalized in the presence of anti-Brucella antibodies. These 261
results demonstrate that the surface N-formyl-perosamine polysaccharides were the moieties 262
responsible of the B. abortus camouflage for opsonization. 263
264
B. abortus is resistant to the bactericidal action of mouse immune serum and PMNs. 265
Since it has been shown that B. abortus ∆wadC and ∆per mutants are more sensitive to the 266
bactericidal action of bovine serum than the parental strain (14), then we asked the question, 267
whether murine serum components and PMNs were capable to kill intracellular Brucella. 268
The ∆wadC mutant was resistant to the action of normal mouse serum, as the virulent B. 269
abortus 2308 (Fig. 5C). Moreover, the presence of anti-Brucella antibodies did not have a 270
significant effect on the viability of the bacterium (Fig. 5C). Although serum components 271
opsonized rough ∆per, no bactericidal activity was recorded by mouse serum in the absence 272
or presence of PMNs (Fig. 5D). 273
274
N-formyl-perosamine homopolysaccharides hamper the binding of murine heat-labile 275
serum components. 276
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
13
Since bacterial opsonization commonly occurs via activation and binding of various heat-277
labile serum factors such as complement and fibronectin (29), then we asked the question 278
whether B. abortus was capable to interact with these components. As presented in Fig. 6A, 279
B. abortus was capable to adsorb a significant number of human serum proteins, including 280
complement components. Under the same experimental conditions mouse serum 281
constituents were barely adsorbed by smooth or rough B. abortus strains (Fig. 6A). The low 282
amounts of mouse serum proteins adsorbed by B. abortus strains revealed as faint bands, 283
corresponded to serum albumin, platelet factor 4 and mannose binding protein. With the 284
exception of the latter protein, no other opsonins or complement-related proteins were 285
detected by this proteomic approach. 286
In spite of the overall lack of affinity of the Brucella cell envelope for mouse serum factors, 287
it was revealed by Western blot that the ∆per mutant adsorbed heat-labile complement and 288
fibronectin in higher amounts than the smooth counterpart strain 2308 (Fig. 6B). This was 289
consistent with the phagocytosis of B. abortus ∆per by murine PMNs in the presence of 290
normal but not inactivated serum (Fig. 5A). These results demonstrate that phagocytosis of 291
the ∆per strain by murine PMNs occurs through opsonization of small amounts of heat-292
labile serum components such as complement and fibronectin and that the O chain and NH 293
polysaccharides hamper the accessibility of these components to surface bacterial 294
molecules. 295
296
Intracellular B. abortus induces the cell death of mouse PMNs. 297
We have demonstrated that B. abortus induces de premature cell death of human PMNs (6). 298
Therefore, we explored the effect of both antibody opsonized smooth B. abortus 2308 and 299
the rough mutant B. abortus ∆per on mouse PMNs. As shown in Figure 7, both internalized 300
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
14
bacteria readily induce the cell death of murine PMNs, paralleling the phenomenon 301
observed with human PMNs. 302
303
DISCUSSION 304
305
It has been shown that naïve human, guinea pig, bovine, rat, caprine and canine PMNs 306
readily phagocytize both smooth and rough Brucella species (3-5, 24-26, 30). Moreover, 307
PMNs from some of these animals (humans, guinea pigs, rats and cows) are capable to 308
ingest Brucella in the presence of inactivated normal sera or even in the absence of sera (3, 309
4, 6). In general, the overall behavior of human PMNs is rather similar to that observed 310
with bovine PMNs (4); a phenomenon that is commensurate with the co-evolution of B. 311
abortus and its host. 312
In this work we have shown that that mouse normal serum components do not opsonize 313
smooth B. abortus and that mouse PMNs, in the absence of specific antibodies, do not 314
phagocytize this bacterium. This phenomenon is specific and the molecules responsible for 315
the Brucella camouflage are N-formyl-perosamine surface homopolysaccharides (Fig. 4). 316
This is relevant, since in addition to the hindrance function, these perosamine 317
polysaccharides display other biological properties related to virulence, such as 318
dimerization of MHC class II, blocking of MHC class II antigen presentation and protection 319
against a collection of bactericidal substances (16, 31-33). 320
The absence of NH and O chain polysaccharides uncovers potential targets in the Brucella 321
cell envelope, such as outer membrane proteins, phospholipids, as well as Kdo and lipid A 322
phosphate groups, present in the innermost sections of core moiety (14, 34). Still, mouse 323
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
15
PMNs fail to ingest rough ∆per B. abortus mutant in the absence of classical heat sensitive 324
serum opsonins, such as complement and fibronectin. This is relevant, since mouse PMNs 325
are capable to ingest latex beads and other bacteria such as S. aureus and E. coli, in the 326
presence or absence of complement; albeit the phagocytosis is more efficient in the former 327
case (22, 23). In the absence of opsonization by specific antibodies or complement, PMNs 328
may also ingest microorganisms through the concourse of other PMN-PRRs such as β2 329
integrins, C-type lectins, scavenger or pentraxins receptors, among several (22, 35, 36). 330
This endorses the failure of mouse PMN-PRRs to recognize putative B. abortus PAMPs, 331
such as outer membrane lipoproteins, adhesion-like proteins, ornithine containing lipids, 332
flagellar structures, and phospholipids, among the most conspicuous elements on the 333
surface of brucellae, which in other bacteria are constitutive PAMPs an targets for 334
recognition (7, 8, 14, 27, 37). 335
Although Brucella organisms are more resistant than other Gram negatives to the killing 336
action of human serum and PMNs, still these components are able to kill about 20-30% of 337
virulent smooth B. abortus after 90 minutes (3-5, 38). This bactericidal activity is even 338
more conspicuous and efficient in the case of rough B. abortus (3). In contrast, both smooth 339
and rough B. abortus were totally resistant to the killing actions of mouse PMNs and 340
complement, even in the presence of antibodies. 341
Therefore, it seems that the lack of B. abortus recognition and the failure to kill this 342
bacterium, works in at least three different levels of the innate immune system: i) the 343
absence of binding of natural opsonins to the surface of smooth Brucella; ii) the lack of 344
recognition of putative PAMPs by murine PMN-PRRs, and; iii) the virtually absolute 345
resistance of Brucella to the killing action of mouse complement ‒either the alternative or 346
classical pathways‒ and PMNs. 347
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
16
Regarding the interaction of Brucella with human complement, it has been demonstrated 348
that these bacteria are considerably more resistant to the killing action of human serum than 349
other Gram negatives (5, 39). In spite of this, the binding of human complement 350
components to the surface of B. abortus was conspicuous and commensurate with the 351
opsonization of this bacterium by normal serum. This is in clear contrast to the absence of 352
opsonization and brucellicidal activity observed by mouse serum. This is striking, since it 353
has been proposed that bacterial opsonization by mouse complement is similar to that of 354
human complement and that mouse has significant quantities of complement activity and 355
high serum levels of C3 as well as other complement proteins (40). However, genetic and 356
structural differences have been demonstrated between human and mouse complement C3 357
and C4 components (41). For instance while human C3 is readily inhibited by compstatin, 358
mouse is resistant to this compound due to structural differences in amino acid residues 359
329-534 (42). Likewise mouse C4 does not form the classical C5 convertase activity, due to 360
differences with the human beta-chain segments of C4 and other regions of the molecule 361
contributing to C5 binding (43). 362
Although the differences between murine and human surface PMN-PRRs have not been 363
explored in detail, there are some discrete features that may be relevant. While in humans 364
the complement receptor proteins CR1 and CR2 are two different molecules; in mouse 365
PMNs, they constitute a single “chimeric” CR1/CR2 molecule with different affinities for 366
various complement components (44). In addition, the affinity of the G-protein coupled 367
receptor for the chemoattractant and activator fMLP is lower in mouse than in human 368
PMNs (45). Likewise, while the human PMN receptor CXCR1 uses as substrate CXCL8, 369
the mouse counterpart uses CXCL6 (46). Finally, some receptors present in mouse PMNs 370
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
17
(e.g. the sialic acid receptor CD33, FcγRIIIA, FcγRIII, FcγRIV) are absent in humans, and 371
vice versa (e.g. TLR10, L-selectin binding to E-selectin, FcγRIIA) (47-52). 372
Finally, the almost absolute resistance to the killing action of mouse complement and 373
PMNs, is linked to the absence of activation of these elements by the putative Brucella 374
PAMPs (5, 7, 8, 14, 27, 37), as well as to the bacterial resistance to the microbicidal 375
substances of PMNs (33, 34). Regarding this, lesser lytic activity of mouse complement in 376
comparison to the human counterpart has been described (53). Concerning PMNs, there are 377
several differences in microbicidal components between human and mouse. For instance, 378
mouse PMNs lack defensins (54) and the function of several proteases and ROS activation 379
differ between PMNs of mice and humans (55-57). Whether or not some of these 380
differences are related to the absence of Brucella recognition and killing by murine serum 381
components and PMNs, remain to be investigated. 382
We have previously shown that Brucella induces premature death of human PMNs (6). 383
Commensurate with this, B. abortus also induces the death of these leukocytes once the 384
bacterium has been internalized via Fc receptors, or in the case of the ∆per mutant strain, via 385
other serum opsonins. As proposed (6), this phenomenon may favor the non-phlogistic 386
removal of infected dying PMNs by Mϕ and DCs, favoring the dispersion of Brucella in the 387
organism, following a “Trojan horse” effect. In this sense, the premature death of mouse 388
PMNs parallels those of humans and therefore it is a useful model to explore this hypothesis. 389
The mouse model has provided a substantial body of information concerning the 390
pathobiology and immunology of brucellosis (12, 13). Still, our approach has revealed 391
significant differences between front-line elements of innate immunity between mice and 392
other hosts that may have profound influence in downstream mechanisms of the immune 393
response. This is not trivial, since the outcome of brucellosis in a given animal species may 394
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
18
be determined during the initial stages of the immune recognition. Therefore, the 395
differences between the immune system of mice and other mammals (58) should prevent us 396
from making direct extrapolations and encourage us to dissect the mechanisms behind 397
them. 398
399
ACKNOWLEDGMENTS 400
401
We thank the research teams of PIET of the Universidad Nacional, CIET of the Universidad 402
de Costa Rica, Ignacio Moriyón and Raquel Conde (University of Navarra, Pamplona, 403
Spain), for providing LPS samples, wadC and per constructs, Alexandra Rucavado (ICP, 404
University of Costa Rica) for the anti-fibronectin antibody and fibronectin control, and 405
Caterina Guzmán-Verri for her helpful discussions. 406
This work was partially supported by grants Fondo Especial de la Educación Superior 407
(FEES-CONARE) de Costa Rica, projects UNA-SIA-0505-13, UNA-SIA-0504-13 (UCR-408
803-B4-654), UNA-SIA-0248-13 UNA-SIA- 0434-14 (UCR-803-B5-653). Red Temática de 409
Brucelosis, Universidad de Costa Rica, project 803-B3-761, Consejo Nacional para 410
Investigaciones Científicas y Tecnológicas (CONICIT-FORINVES) grant FV-0004-13. The 411
International Center for Genetic Engineering and Biotechnology, grant CRP/12/007. 412
413
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
19
REFERENCES 414
415
1. Moreno E, Moriyón I. 2006. The genus Brucella, p 315–456. In Dworkin M, Falkow S, 416
Rosenberg E, Schleifer K-H, Stackebrant E (ed), The prokaryotes, vol 5. Springer 417
Verlag, New York, NY. 418
2. Ackermann MR, Cheville NF, Deyoe BL. 1988. Bovine ileal dome lymphoepithelial 419
cells: endocytosis and transport of Brucella abortus strain 19. Vet Pathol 25: 28-35. 420
3. Kreutzer DL, Dreyfus LA, Robertson DC. 1979. Interaction of polymorphonuclear 421
leukocytes with smooth and rough strains of Brucella abortus. Infect Immun 23: 737-422
742. 423
4. Riley LK, Robertson DC. 1984. Ingestion and intracellular survival of Brucella abortus 424
in human and bovine polymorphonuclear leukocytes. Infect Immun 46: 224-230. 425
5. Barquero-Calvo E, Chaves-Olarte E, Weiss DS, Guzmán-Verri C, Chacón-Díaz C, 426
Rucavado A, Moriyón I, Moreno E. 2007.Brucella abortus uses a stealthy strategy to 427
avoid activation of the innate immune system during the onset of infection. PLoS One 2: 428
e631. 429
6. Barquero-Calvo E, Mora-Cartín R, Arce-Gorvel V, de Diego J,L, Chacón-Díaz C, 430
Chaves-Olarte E, Guzmán-Verri C, Buret AG, Gorvel JP, Moreno E. 2015. Brucella 431
abortus induces the premature death of human neutrophils through the action of its 432
lipopolysaccharide. PLoS Pathog 11: e1004853. 433
7. Barquero-Calvo E, Conde-Álvarez R, Chacón-Díaz C, Quesada-Lobo L, 434
Martirosyan A, Guzmán-Verri C, Iriarte M, Mancek-Keber M, Jerala R, Gorvel 435
JP, Moriyón I, Moreno E, Chaves-Olarte E. 2009.The differential interaction of 436
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
20
Brucella and Ochrobactrum with innate immunity reveals traits related to the evolution 437
of stealthy pathogens. PLoS One 4: e5893. 438
8. Palacios-Chaves L, Conde-Álvarez R, Gil-Ramírez Y Zúñiga-Ripa A, Barquero-439
Calvo E, Chacón-Díaz C, Chaves-Olarte E, Arce-Gorvel V, Gorvel JP, Moreno E, 440
de Miguel MJ, Grilló MJ, Moriyón I, Iriarte M. 2011. Brucella abortus ornithine 441
lipids are dispensable outer membrane components devoid of a marked pathogen-442
associated molecular pattern. PLoS One 6: e16030. 443
9. von Bargen K, Gagnaire A, Arce-Gorvel V, de Bovis B, Baudimont F, Chasson L, 444
Bosilkovski M, Papadopoulos A, Martirosyan A, Henri S, Mège JL, Malissen B, 445
Gorvel JP. 2015. Cervical lymph nodes as a selective niche for Brucella during oral 446
infections. PLoS ONE 10: e0121790. 447
10. Copin R, Vitry MA, Hanot-Mambres D, Machelart A, De Trez C, Verwinden JM, 448
Magez S, Akira S, Ryffel B, Carlier Y, Letesson JJ, Muraille E. 2012. In situ 449
microscopy analysis reveals local innate immune response developed around Brucella 450
infected cells in resistant and susceptible mice. PLoS Pathog 8: e1002575. 451
11. Barquero-Calvo E, Martirosyan A, Ordoñez-Rueda D, Arce-Gorvel V, Alfaro-452
Alarcón A, Lepidi H, Malissen B, Malissen M, Gorvel JP, Moreno E. 2013. 453
Neutrophils exert a suppressive effect on Th1 responses to intracellular pathogen 454
Brucella abortus. PLoS Pathog 9: e1003167. 455
12. Grilló MJ, Blasco JM, Gorvel JP Moriyón I, Moreno E. 2012. What have we learned 456
from brucellosis in the mouse model? Vet Res 43: 29. 457
13. Silva TM, Costa EA, Paixão TA, Tsolis RM, Santos RL. 2011. Laboratory animal 458
models for brucellosis research. J Biomed Biotechnol 2011: ID518323. 459
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
21
14. Conde-Álvarez R, Arce-Gorvel V, Iriarte M, Manček-Keber M, Barquero-Calvo 460
E, Palacios-Chaves, L, Chacón-Díaz C, Chaves-Olarte E, Martirosyan A, von 461
Bargen K, Grilló MJ, Jerala R, Brandenburg K, Llobet E, Bengoechea JA, Moreno 462
E, Moriyón I, Gorvel JP. 2012. The lipopolysaccharide core of Brucella abortus acts as 463
a shield against innate immunity recognition. PLoS Pathog 8: e1002675. 464
15. Chacón-Díaz C, Muñoz-Rodríguez M, Barquero-Calvo E, Guzmán-Verri C, 465
Chaves-Olarte E, Grilló MJ, Moreno E. 2011. The use of green fluorescent protein as 466
a marker for Brucella vaccines. Vaccine 29: 577-582. 467
16. Aragón V, Díaz, R, Moreno E, Moriyón I. 1996. Characterization of Brucella abortus 468
and Brucella melitensis native haptens as outer membrane O-type polysaccharides 469
independent from the smooth lipopolysaccharide. J Bacteriol 178: 1070-1079. 470
17. Moreno E, Stackebrandt E, Dorsch M, Wolters J, Busch M, Mayer H. 1990. 471
Brucella abortus 16S rRNA and lipid A reveal a phylogenetic relationship with 472
members of the alpha-2 subdivision of the class Proteobacteria. J Bacteriol 172: 3569-473
3576. 474
18. McKinney MM, Parkinson A. 1987. A simple, non-chromatographic procedure to 475
purify immunoglobulins from serum and ascites fluid. J Immunol Methods 96: 271-278. 476
19. Boxio R, Bossenmeyer-Pourié C, Steinckwich N, Dournon C, Nüsse O. 2004. Mouse 477
bone marrow contains large numbers of functionally competent neutrophils. J Leukoc 478
Biol 75: 604-611. 479
20. Chacón-Díaz C, Altamirano-Silva P, González-Espinoza G, Medina M-C, Alfaro-480
Alarcón A, Bouza-Mora L, Wong M, Barquero-Calvo E, Rojas N, Guzmán-Verri 481
C, Moreno E, Chaves-Olarte E. 2015. Brucella canis is an intracellular pathogen 482
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
22
inducing a lower proinflammatory response than smooth zoonotic counterparts. Infect 483
Immun 83: 4861-4870. 484
21. Lomonte B, Tsai, WC, Ureña-Díaz JM, Sanz L, Mora-Obando D, Sánchez EE, Fry 485
BG, Gutiérrez JM, Gibbs HL, Calvete JJ. 2014. Venomics of New World pit vipers: 486
genus-wide comparisons of venom proteomes across Agkistrodon. J Proteomics 96: 103-487
116. 488
22. Anderson KE, Boyle KB, Davidson K, Chessa TA, Kulkarni S, Jarvis GE, 489
Sindrilaru A, Scharffetter-Kochanek K, Rausch O, Stephens LR, Hawkins PT. 490
2008. CD18-dependent activation of the neutrophil NADPH oxidase during 491
phagocytosis of Escherichia coli or Staphylococcus aureus is regulated by class III but 492
not class I or II PI3Ks. Blood 112: 5202-5211. 493
23. Hart PH, Spencer LK, Nikoloutsopoulos A, Lopez AF, Vadas MA, McDonald PJ, 494
Finlay-Jones JJ. 1986. Role of cell surface receptors in the regulation of intracellular 495
killing of bacteria by murine peritoneal exudate neutrophils. Infect Immun 52: 245-251 496
24. Canning PC, Deyoe BL, Roth JA. 1988. Opsonin-dependent stimulation of bovine 497
neutrophil oxidative metabolism by Brucella abortus. Am J Vet Res 49: 160-163. 498
25. Meador VP, Deyoe BL, Cheville NF. 1989. Pathogenesis of Brucella abortus infection 499
of the mammary gland and supramammary lymph node of the goat. Vet Pathol 26: 357-500
368. 501
26. Victor J, Pollack AD, Raymond R, Valliant JR. 1952. Studies on phagocytosis 502
determination of blood opsonin for Brucella. J Bacteriol 64: 121-130. 503
27. Moreno E, Kurtz, RS, Berman DT. 1984. Induction of immune and adjuvant 504
immunoglobulin G responses in mice by Brucella lipopolysaccharide. Infect Immun 46: 505
74-80. 506
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
23
28. Bundle DR, Cherwonogrodzky JW, Gidney MA Meikle PJ, Perry MB, Peters T. 507
1989. Definition of Brucella A and M epitopes by monoclonal typing reagents and 508
synthetic oligosaccharides. Infect Immun 57: 2829-2836. 509
29. Vuento M, Salonen E, Salminen K, Pasanen M, Stenman UK. 1980. 510
Immunochemical characterization of human plasma fibronectin. Biochem J 191: 719-511
727. 512
30. Gallego MC, Lapeña MA. 1990. The interaction of Brucella melitensis 16-M and 513
caprine polymorphonuclear leukocytes. Comp Immunol Microbiol Infect Dis 13: 59-65. 514
31. Escola JM, Moreno E, Chavrier P, Gorvel JP. 1994. The O-chain of Brucella 515
abortus lipopolysaccharide induces SDS-resistant MHC class II molecules in mouse B 516
cells. Biochem Biophys Res Commun 203: 1230-1236. 517
32. Forestier C, Deleuil F, Lapaque N, Moreno E, Gorvel JP. 2000. Brucella abortus 518
lipopolysaccharide in murine peritoneal macrophages acts as a down-regulator of T cell 519
activation. J Immunol 165: 5202-5210. 520
33. Freer E, Moreno E, Moriyón I, Pizarro-Cerdá J, Weintraub A, Gorvel JP. 1996. 521
Brucella-Salmonella lipopolysaccharide chimeras are less permeable to hydrophobic 522
probes and more sensitive to cationic peptides and EDTA than are their native Brucella 523
sp. counterparts. J Bacteriol 178: 5867-5876. 524
34. Martínez de-Tejada G, Pizarro-Cerdá J, Moreno E, Moriyón I. 1995. The outer 525
membranes of Brucella spp. are resistant to bactericidal cationic peptides. Infect Immun 526
63: 3054-3061. 527
35. Bottazzi B, Doni A, Garla C, Mantovani A. 2010. An integrated view of humoral 528
innate immunity: pentraxins as a paradigm. Annu Rev Immunol 28: 157-183. 529
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
24
36. Futosi K, Fodor S, Mócsai A. 2013. Neutrophil cell surface receptors and their 530
intracellular signal transduction pathways. Int Immunopharmacol 17: 638-650. 531
37. Martirosyan A, Moreno E, Gorvel JP. 2011. An evolutionary strategy for a stealthy 532
intracellular Brucella pathogen. Immunol Rev 240: 211-234. 533
38. Fernández-Prada CM, Zelazowska EB, Nikolich M, Hadfield TL, Roop RM 2nd, 534
Robertson GL, Hoover DL. 2003. Interactions between Brucella melitensis and human 535
phagocytes: bacterial surface O-Polysaccharide inhibits phagocytosis, bacterial killing, 536
and subsequent host cell apoptosis. Infect Immun 71: 2110-2119. 537
38. Fernandez-Prada CM1, Nikolich M, Vemulapalli R, Sriranganathan N, Boyle 538
SM, Schurig GG, Hadfield TL, Hoover DL. 2001. Deletion of wboA enhances 539
activation of the lectin pathway of complement in Brucella abortus and Brucella 540
melitensis. Infect Immun 69: 4407-4416. 541
40. Osmers I, Szalai AJ, Tenner AJ, Barnum SR. 2006. Complement in BuB/BnJ mice 542
revisited: serum C3 levels and complement opsonic activity are not elevated. Mol 543
Immunol 43: 1722-1725. 544
41. Odink KG, Fey G, Wiebauer K, Diggelmann H. 1981. Mouse complement 545
components C3 and C4. Characterization of their messenger RNA and molecular cloning 546
of complementary DNA for C3. J Biol Chem 256: 1453-1458. 547
42. Tamamis P1, Pierou P, Mytidou C, Floudas CA, Morikis D, Archontis G. 2011. 548
Design of a modified mouse protein with ligand binding properties of its human analog 549
by molecular dynamics simulations: the case of C3 inhibition by compstatin. Proteins 550
79: 3166-3179. 551
43. Ebanks RO, Isenman DE. 1996, Mouse complement component C4 is devoid of 552
classical pathway C5 convertase subunit activity. Mol Immunol 33: 297-309. 553
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
25
44. Jacobson AC, Wei, JH. 2008. Comparative functional evolution of human and mouse 554
CR1 and CR2. J Immunol 181: 2953-2959. 555
45. Gao JL, Murphy PM. 1993. Species and subtype variants of the N-formyl peptide 556
chemotactic receptor reveal multiple important functional domains. J Biol Chem 268: 557
25395-25401 558
46. Fan X, Patera AC, Pong-Kennedy A, Deno G, Gonsiorek W, Manfra DJ, 559
Vassileva G, Zeng M, Jackson C, Sullivan L, Sharif-Rodriguez W, Opdenakker G, 560
Van Damme J, Hedrick JA, Lundell D, Lira SA, Hipkin RW. 2006. Murine CXCR1 561
is a functional receptor for GCP-2/CXCL6 and interleukin-8/CXCL8. J Biol Chem 282: 562
11658-11666. 563
47. Brinkman-Van der Linden EC, Angata T, Reynolds SA, Powell LD, Hedrick 564
S.M., Varki A. 2003.CD33/Siglec-3 binding specificity, expression pattern, and 565
consequences of gene deletion in mice. Mol Cell Biol 23: 4199-4206. 566
48. Bruhns P. 2012. Properties of mouse and human IgG receptors and their contribution 567
to disease models. Blood 119:5640-5649. 568
49. Hasan U, Chaffois C, Gaillard C, Saulnie, V, Merck E, Tancredi S, Guiet C, 569
Brière F, Vlach J, Lebecque S, Trinchieri G, Bates EE. 2005. Human TLR10 is a 570
functional receptor, expressed by B cells and plasmacytoid dendritic cells, which 571
activates gene transcription through MyD88. J Immunol 174: 2942-2950. 572
50. Jönsson F, Mancardi DA, Zhao W, Kita Y, Iannascoli B, Khun H, van Rooijen N, 573
Shimizu T, Schwartz LB, Daëron M, Bruhns P. 2012. Human FcγRIIA induces 574
anaphylactic and allergic reactions. Blood 119:2533-2544. 575
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
26
51. Marois L, Paré G, Vaillancourt M, Rollet-Labelle E, Naccache PH. 2011. Fc 576
gammaRIIIb triggers raft-dependent calcium influx in IgG-mediated responses in 577
human neutrophils. J Biol Chem 286:3509-3519. 578
52. Zöllner O, Lenter MC, Blanks JE, Borges E, Steegmaier M, Zerwes HG, 579
Vestweber D. 1997. L-selectin from human, but not from mouse neutrophils binds 580
directly to E-selectin. J Cell Biol 136:707-716. 581
53. Ish C, Ong GL, Desai N, Mattes MJ. 1993. The specificity of alternative complement 582
pathway-mediated lysis of erythrocytes: a survey of complement and target cells from 25 583
species. Scand J Immunol. 38: 113-122. 584
54. Eisenhauer PB, Lehrer RI. 1992. Mouse neutrophils lack defensins. Infect Immun 585
60: 3446-3447. 586
55. Bagaitkar J, Matute JD, Austin A, Arias AA, Dinauer MC. 2012. Activation of 587
neutrophil respiratory burst by fungal particles requires phosphatidylinositol 3-588
phosphate binding to p40phox in humans but not in mice. Blood 120:3385-3387. 589
56. Kalupov T, Brillard-Bourdet M, Dadé S, Serrano H, Wartelle J, Guyot N, Juliano 590
L, Moreau T, Belaaouaj A, Gauthier F. 2009. Structural characterization of mouse 591
neutrophil serine proteases and identification of their substrate specificities:relevance to 592
mouse models of human inflammatory diseases. J Biol Chem 284:34084-34091. 593
57. Wiesner O, Litwiller RD, Hummel AM, Viss MA, McDonald CJ, Jenne DE, Fass 594
DN, Specks U. 2005. Differences between human proteinase 3 and neutrophil elastase 595
and their murine homologues are relevant for murine model experiments. FEBS Lett 596
579:5305-5312. 597
58. Mestas J, Hughes CC. 2004. Of mice and not men:differences between mouse and 598
human immunology. J Immunol 172: 2731-2738. 599
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
27
59. Kubler-Kielb J, Vinogradov E. 2013. The study of the core part and non-repeating 600
elements of the O-antigen of Brucella lipopolysaccharide. Carbohydr Res 366: 33-37. 601
60. Iriarte M, González D, Delrue RM, Monreal D, Conde R, López-Goñi I, Letesson 602
JJ. 2004. Brucella lipopolysaccharide: structure, biosynthesis and genetics, p152-183. In 603
López-Goñi I, Moriyón I (ed), Brucella Molecular and Cellular Biology. Wymondham, 604
Horizon Bioscience, UK. 605
61. Gil-Ramírez Y, Conde-Álvarez R, Palacios-Chaves L, Zúñiga-Ripa A, Grilló MJ, 606
Arce-Gorvel V, Hanniffy S, Moriyón I, Iriarte M. 2014. The identification of wadB, a 607
new glycosyltransferase gene, confirms the branched structure and the role in virulence 608
of the lipopolysaccharide core of Brucella abortus. Microb Pathog 73: 53-59. 609
610
Figure legends 611
612
FIG. 1 B. abortus and its Br-LPS do not induce the cell death of mouse PMNs. (A) 613
Heparinized blood of human or C57BL/6 mice was incubated with B. abortus 2308-GFP at 614
the indicated MOI for two hours and the PMNs population gated and analyzed by cytometry 615
using Annexin V as a cell death marker. (B) Heparinized blood of C57BL/6 mice was 616
incubated with B. abortus 2308-GFP at MOI 10 (green line) MOI 100 (blue line) or PBS 617
(red line) for two hours and PMNs population gated and analyzed by cytometry using 618
Annexin V as a cell death marker. (C) Heparinized blood of human or C57BL/6 mice was 619
incubated with Br-LPS at the indicated concentrations for two hours and PMNs population 620
gated and analyzed by cytometry for cell death markers using AquaDead and Annexin V. 621
The proportions of PMNs positive for the respective marker are presented with each section 622
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
28
of the corresponding cytogram. While human PMNs acquire death cell markers, mouse 623
PMNs remain unaffected. The figure represents one experiment of at least four repetitions. 624
625
FIG. 2 Naïve murine PMN do not phagocytize B. abortus. (A) Blood from the indicated 626
species was incubated with B. abortus-GFP, S. aureus or E. coli at MOI 20 for one hour. 627
Blood smears were then fixed and mounted with ProLong Gold Antifade Reagent with 628
DAPI. At least 50 PMNs were counted per sample and the number of intracellular bacterial 629
in each PMN and the proportion of phagocytosis estimated under fluorescent microscopy. 630
(B) Comparison between (a) C57BL/6 mouse PMNs incubated with B. abortus 2308-RFP at 631
MOI 100; (b) human PMNs incubated with B. abortus-GFP at MOI 50; (c) Giemsa staining 632
of C57BL/6 mouse PMNs incubated with latex beads at MOI 100. Magnification 200 × (a 633
panel) and 400 × (b and c right panels). Images were cut from the microscope field, 634
contrasted and saturated using Hue tool to obtain suitable color separation. Similar results 635
were obtained with Balb/c mice. 636
637
FIG. 3 Mouse PMNs phagocytize B. abortus after antibodies are generated. CD-1 mice 638
were infected intraperitoneally with 106 CFU; then, mouse blood was collected at different 639
days of infection and incubated with B. abortus 2308-RFP (MOI 50) for one hour. Blood 640
smears were then fixed and mounted with ProLong Gold Antifade Reagent with DAPI. At 641
least 50 PMNs were counted per sample and the number of intracellular bacterial in each 642
PMN determined. Minus (-) and plus (+) signs on top of the bars indicate the agglutination 643
titer for the day evaluated after intraperitoneal injection with B. abortus. Two positive marks 644
indicate agglutination in a serum dilution of 1/125. (B) Mouse PMNs of CD-1 mice were 645
incubated with B. abortus 2308-RFP at MOI 50 for two hours and under different conditions 646
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
29
of opsonization. Then, the number of phagocytized bacteria was recorded by fluorescent 647
microscopy. (C) CD-1 mouse PMNs incubated with B. abortus 2308-RFP and anti-Brucella 648
mouse serum at MOI 100; magnification 200 ×. Image was cut from the microscope field, 649
contrasted and saturated using Hue tool to obtain suitable color separation. Similar results 650
were obtained with C57BL/6 and Balb/c mice. 651
652
Figure 4. Schematic structure of B. abortus outer membrane showing the Br-LPS and 653
NH. The O-polysaccharide and native hapten (NH) are unbranched linear homopolymers of 654
α-1,2-linked 4,6-dideoxy-4-formamido-D-mannopyranosyl units (N-formyl-perosamine) 655
with an average chain length of 96 to 100 glycosyl subunits (28). While the NH is not 656
directly bound to the lipid membrane, the O-polysaccharide is linked to a core bifurcating 657
oligosaccharide composed of βGlcN-6-βGlcN-4-βGlcN(-6-βGlcN)-3-αMan(-6-αGlc)-5-658
KDO1(-1-KDO1)-Lipid-A immersed in the outer membrane. Branching from KDO2 is 659
αPerNFo-(-2PerNFo)n-2PerNF-2-αMan-3-αMan-3-βQuiNAc-4-βGlc-4-KDO2-4-KDO2 660
(59). The KDO1 is linked to the lipid A composed of a backbone of diaminoglucose (DAG) 661
disaccharide, substituted with phosphates (P) and amide and ester-linked long chain 662
saturated (C16:0 to C18:0) and hydroxylated (3-OH-C12:0 to 29-OH-C30:0) fatty acids (17, 60). 663
The lipid A is bound to the outer membrane. Ketodeoxyoctulosonic acid (KDO), mannose 664
(Man), Acetyl-quinovosamine (QuiN), glucose (Glc). The ∆wadC mutation precludes the 665
incorporation of the βGlcN-6-βGlcN-4-βGlcN(-6-βGlcN)-3-αMan(-6-αGlc)- 666
oligosaccharide to the KDO1 (marked by a gray area) while the ∆per mutation precludes the 667
incorporation of αPerNFo-(-2PerNFo)n-2PerNF- homopolymer (61). 668
669
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
30
FIG. 5 Surface N-formyl-perosamine polysaccharides are responsible for Brucella 670
camouflage. (A) Smooth B. abortus ∆wadC and rough ∆per mutants were incubated with 671
Balb/c bone marrow PMNs at the respective MOI in the presence of normal or inactivated 672
mouse serum and the phagocytosis recorded as in FIG. 2. The insert corresponds to a PMN 673
with internalized B. abortus ∆per (MOI 50): the infected cell was fixed and mounted with 674
ProLong Gold Antifade Reagent with DAPI and observed under the fluorescent microscope 675
at magnification 200 ×; the Image was cut from the microscope field, contrasted and 676
saturated using Hue tool to obtain suitable color separation. (B) Y. enterocolitica O.9 (MOI 677
100) was incubated with Balb/c bone marrow PMNs in the presence of normal serum or 678
serum with anti-Brucella antibody (2%) and the phagocytosis recorded. (C) B. abortus 2308 679
and the ∆wadC mutant (CFU 106) were incubated with normal, inactivated or normal mouse 680
serum containing antibodies against Brucella and the bacterial viability recorded after 45 681
minutes hours of incubation. (D) B. abortus ∆per was incubated in the presence of normal or 682
inactivated mouse serum or with Balb/c PMNs in normal serum and the bacterial viability 683
estimated at indicated times. 684
685
FIG. 6 Normal mouse serum proteins do not opsonize smooth B. abortus. (A) 686
Commassie blue stained 10% SDS-PAGE of mouse (M) and human (H) proteins (10 687
μg/well) eluted from the surface of B. abortus 2308 or ∆per after incubation of the 688
respective serum with viable bacterial cells. Individual lines, corresponding to the different 689
eluted fractions, were cut out and separated from the main gel as indicated in the figure; then 690
each gel line was horizontally sliced (about 2 mm slices, from top to bottom) and identity of 691
proteins in each slice determined by proteomic analysis. (B) Western blotting of mouse 692
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from
31
proteins (10 μg/well) eluted from the surface of B. abortus 2308 or ∆per after incubation of 693
the respective serum with viable bacterial cells. Western blots were developed with 694
monoclonal antibodies against mouse C´3 or monospecific rabbit IgG anti-fibronectin. 695
696
FIG. 7 Induction of mouse PMNs cell death after phagocytosis of antibody- opsonized 697
smooth B. abortus 2308 or rough ∆per mutant. (A) Heparinized blood of C57BL/6 mice 698
was incubated with B. abortus 2308-GFP or with B. abortus 2308-GFP opsonized with 699
antibody against Br-LPS at the indicated MOIs for four hours and then the PMNs population 700
gated and analyzed by cytometry using Annexin V as a cell death marker. Negative controls 701
were either treated with PBS or antibody alone. Similar results were observed with bone 702
marrow derived PMNs (B) Bone marrow PMNs of Balb/c mice were incubated with rough 703
mutant B. abortus ∆per at MOI 50 or PBS for four hours and PMNs population gated and 704
analyzed by cytometry using Annexin V as a cell death marker. The figure represents one 705
experiment of at least four repetitions. 706
on March 6, 2018 by guest
http://iai.asm.org/
Dow
nloaded from