Previous Article | Next Article ![]()
Infection and Immunity, February 2003, p. 621-628, Vol. 71, No. 2
0019-9567/03/$08.00+0 DOI: 10.1128/IAI.71.2.621-628.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
Pathologie Infectieuse et Immunologie,1 Service commun de cytométrie en flux, Institut National de la Recherche Agronomique, Nouzilly,3 Institut National de la Santé et de la Recherche Médicale U-431, Montpellier, France2
Received 8 July 2002/ Returned for modification 3 September 2002/ Accepted 1 November 2002
|
|
|---|
), interleukin-12 p40 (IL-12p40), gamma interferon (IFN-
), and IL-10 mRNAs were similarly induced in spleens of the two strains. In contrast, the mRNA of KC, a C-X-C chemokine, was induced only in infected C.CB mice at this time. This pattern of mRNA expression was maintained at 14 days p.i., with IFN-
and IL-12p40 mRNAs being more intensively induced in the infected C.CB mice, but TNF-
mRNA was no longer induced. The higher recruitment of neutrophils observed in the spleens of infected C.CB mice could explain the temporary susceptibility of C.CB mice to B. melitensis infection. In contrast to infections with Salmonella, Leishmania, and Mycobacterium, the expression of the Nramp1 gene appears to be of limited importance for the natural resistance of mice to Brucella. |
|
|---|
The bovine homolog of the murine Nramp1 gene (known as bovine NRAMP1) has been cloned, and the cDNA has been sequenced (16). Bovine NRAMP1 is expressed primarily in the macrophages of the reticuloendothelial system. An in vitro model for the expression of resistance- and susceptibility-associated alleles of bovine NRAMP1 as stable transgenes in the murine RAW264.7 macrophage cell line was developed recently (3). The point mutation within the coding region of murine Nramp1 that distinguishes between resistant and susceptible alleles of the mouse Bcg gene (43) has never been documented in cattle or the other mammalian species in which the Nramp1 homolog has been sequenced. Inbred mouse strains differ in their ability to control B. abortus infection: C57BL/10 mice more effectively control the initial replication of the virulent B. abortus strain 2308 than do BALB/c mice, and they clear the infection within 16 weeks whereas BALB/c mice remain infected (24). A recent study using macrophages from these two strains of mice did not find any difference in the ability of these macrophages to control the intracellular growth of B. abortus strain 2308 (39). Both resistant C57BL/10 and susceptible BALB/c mouse strains are homozygous for the Nramp1s allele, and susceptibility to B. abortus infection has been reported in the CBA mouse strain, which is homozygous for the Nramp1r allele (24). Since the genetic control of B. abortus infection is known to be a multigenic trait, we compared the ability of two strains of mice, BALB/c (Nramp1s) and the congenic C.CB mouse (Nramp1r), to control B. melitensis infection in order to study the effect of the Nramp1 gene in the context of a common genetic BALB/c background. The effect of Nramp1 expression on the intracellular survival of B. melitensis in macrophages and on the inflammatory response induced by B. melitensis infection was also investigated.
|
|
|---|
Bacteria and infection. Mice were infected by the intraperitoneal (i.p.) route with 104 CFU of the B. melitensis H38S virulent strain. This dose induces splenic infection that persists for at least 10 weeks (8). This strain was grown for 24 h on Trypticase soy agar (Life Technologies, Cergy Pontoise, France) supplemented with 0.1% (wt/vol) yeast extract (Difco, Detroit, Mich.) (TSA-YE), harvested in buffered saline solution, adjusted spectrophotometrically at 600 nm, and diluted to 5 x 104 CFU ml-1. Viable counts were determined retrospectively by counting on TSA-YE plates. Eight infected mice from the BALB/c or C.CB strains were killed by cervical dislocation on days 2, 4, 6, 14, 21, and 28 postinfection (p.i.). Their spleens and livers were harvested, weighed, and frozen at -20°C. After the tissues were homogenized in BSS, bacterial counts were performed on TSA-YE plates. To normalize the distribution of individual counts for statistical analysis, the number of CFU per organ was transformed as log10 (CFU log10/CFU), as previously reported (7).
Gentamicin protection assay. Infected BALB/c and C.CB mice were killed on days 5 and 14 p.i. The spleens were removed, and individual cell suspensions (four mice per strain per time point) were prepared by the conventional method of pressing the tissues through nylon gauze. Erythrocytes were lysed by adding 0.155 M NH4Cl to the cell pellet, and the cells were washed and finally adjusted to 107 cells per ml of gentamicin-free minimum essential medium (MEM) (Life Technologies) supplemented with 5% fetal calf serum (FCS) (Life Technologies). A 1-ml volume of each cell suspension was collected, the cells were lysed in 0.1% Triton X-100 and the viable cells were counted by plating serial dilutions on TSA-YE. The cell suspensions were then centrifuged and resuspended at 107 cells per ml of MEM-5% FCS supplemented with gentamicin (200 µg ml-1) (Sigma, St. Louis, Mo.) to kill any extracellular or adherent Brucella organisms. The cells were incubated with stirring at room temperature for 60 min. They were washed twice with gentamicin-free medium and adjusted to 107 cells per ml of gentamicin-free medium, and counts of the viable Brucella organisms were done again on 1 ml of this cell suspension plated on TSA-YE. The counts were expressed as log10 CFU per spleen with or without incubation with gentamicin.
Intracellular survival assay from macrophages infected in vivo by B. melitensis. BALB/c and C.CB mice were infected i.p., as described above, with 104 CFU of H38S strain and killed 3 days p.i. Peritoneal exudate cells (PEC) were harvested by washing the peritoneal cavity with 3 ml of MEM-FCS. PEC from three mice per strain were pooled, centrifuged at 600 x g for 10 min, and then resuspended in RPMI 1640 medium plus 20 mM HEPES (Life Technologies) supplemented with 10% FCS, 2 mM L-glutamine, 1 mM sodium pyruvate, 5 x 10-5 M ß-mercaptoethanol, and 100 µg of gentamicin ml-1 to kill any extracellular Brucella organisms. The cells were adjusted to 5 x 105 cells ml-1, seeded in 24-well plates (1 ml per well), and incubated at 37°C under 5% CO2. At various time intervals, the viable counts of intracellular Brucella organisms were determined after the cells were washed with gentamicin-free medium, and then the cells were lysed with 0.1% Triton X-100. The results are expressed as mean of log10 CFU per well ± standard deviation from three wells at each time point.
In vitro intracellular survival assay with Nramp1 RAW 264.7 transfectants.
RAW264.7 transfectant cell lines R21 (Nramp1 deficient) and R37 (Nramp1 expressing) were provided by C. H. Barton and cultured as described previously (2). Stable Nramp1 expression was confirmed in R37 but not R21 cell lines by Western blotting using anti-Nramp1 antibodies (C. H. Barton, personal communication). Cells were grown without geneticin and seeded at 3 x 105 cells ml-1 in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FCS in 24-well plates and allowed to adhere for 10 h at 37°C under 5% CO2. The cells were then exposed overnight to recombinant murine gamma interferon (IFN-
) (10U/ml) (19301T [Pharmingen, San Diego, Calif.]). The inoculum of B. melitensis H38S was prepared from a 24-h culture on a TSA-YE slope. Bacteria were collected, washed, and resuspended in DMEM-10% FCS to 6 x 107 CFU ml-1. The exact titer of the infectious inoculum was determined by plating on TSA-YE plates. For infection, the cells were incubated with 250 µl of bacteria suspension (corresponding to a multiplicity of infection of 50:1). The plates were centrifuged (3 min at 1,500 x g) to optimize the contact between the cells and bacteria and incubated for a further 60 min at 37°C with 5% CO2. The cells were then washed thoroughly with MEM-FCS to remove any nonadherent bacteria and then reincubated for a further 60 min with 1 ml of complete RPMI medium supplemented with gentamicin (100 µg ml-1) to kill any remaining extracellular bacteria. At this time (T0) and then at 5, 24, and 48 h p.i., the culture supernatant from three wells was removed and the cells were washed with gentamicin-free medium and lysed with 0.1% Triton X-100. Lysates were rapidly diluted in BSS and plated on TSA-YE. The results are expressed as mean log10 CFU per well ± standard deviation from three wells at each time point.
Cell preparation, MAbs, and flow cytometry analysis. Spleen cells were individually prepared from infected and control (uninfected) BALB/c and C.CB mice as described above for the gentamicin protection assay. Cells were finally adjusted to 3 x 107 ml-1 in MEM-FCS with gentamicin (100 µg ml-1). Rat anti-mouse Ly-6G (Gr-1) (RB6-8C5; working dilution, 1:200 [Pharmingen]) and anti-murine pan macrophages (BM8; working dilution, 1:50 [Bachem, Voisins-le-Bretonneux, France]) were used to analyze the neutrophils (18) and major subpopulations of resident tissue macrophages (29), respectively. R-Phycoerythrin (RPE)-conjugated goat anti-rat immunoglobulin G (IgG) F(ab')2 was used as secondary antibody (STAR73; working dilution, 1:50 [Serotec, Kidlington, United Kingdom]). An irrelevant rat IgG2a (MCA1212 [Serotec]) was used as the primary monoclonal antibody (MAb) for a negative control. All antibodies were diluted in 10 mM phosphate-buffered saline (pH 7.4) containing 0.2% (wt/vol) bovine serum albumin (PBS-B) either alone or with 10% mouse serum for RPE anti-rat Ig. For labeling, cells (3 x 106) were incubated at 4°C for 20 min with 50 µl of antibody (primary or secondary) at the appropriate concentration, washed three times with PBS-B, and fixed with 0.5% paraformaldehyde in PBS. The cells were analyzed on a FACStarPlus instrument (Becton Dickinson, San-Jose, Calif.). Two types of gating were performed with forward and side scatter: one for the leukocyte population and the other for phagocyte population. The proportion of nonspecific binding of the secondary antibody was determined using the leukocyte gate and deduced from the proportion of positive cells. Data were expressed as the number of positive cells per spleen by multiplying the proportion of positive cells by the total number of splenocytes per mouse. The proportion of positive cells was also analyzed by gating the phagocyte population, including neutrophils and macrophages.
Analysis of mRNA expression by RT-PCR.
BALB/c and C.CB mice of infected or control groups were killed on days 5 and 14 p.i. Splenocytes from four mice per group were isolated and pooled. Total RNA from 3 x 107 cells was then extracted from each sample by using Trizol (Life Technologies) as described by the manufacturer. The reverse transcription (RT) reaction was then performed at 42°C for 90 min on 20 µg of total RNA, using the Moloney murine leukemia virus reverse transcriptase (Life Technologies) and oligo(dT) (12-18 oligo-dT [Life Technologies]) in the presence of 1 µCi of [32P]dCTP (ICN, Orsay, France), to measure the amount of cDNA synthesized (21). A 1-ng portion (0.2 ng for ß2-microglobulin) of each cDNA was amplified using 2.5 U of Gold Star polymerase (Eurogentec, Seraing, Belgium) and 1 µM specific primers. The specific primers for ß2-microglobulin, interleukin-2 (IL-2), IL-4 (14), tumor necrosis factor alpha (TNF-
) and inducible nitric oxide synthase iNOS (21) and the lengths of the amplicons are described elsewhere. The amplicons were defined as follows: IFN-
(5'-primer, 5'-GCT CTG AGA-CAA-TGA-ACG-CT-3'; 3'-primer, 5'-AAA-GAG-ATA-ATC-TGG-CTC-TGC-3'; 226 bp), IL-12p40 (5'-primer, 5'-CGT-GCT-CAT-GGC-TGG-TGC-AAA-G-3'; 3'-primer, 5'-GCC-CAA-GAA-CTT-GCA-GAT-GAA-G-3'; 316 bp), IL-10 (5'-primer, 5'-ATG-CAG-GAC-TTT-AAG-GGT-TAC-TTG-3'; 3'-primer, 5'-TAG-ACA-CCT-TGG-TCT-TGG-AGC-TTA-3'; 224 bp), IL-18 (5'-primer, 5'-ACT-GTA-CAA-CCG-CAG-TAA-TAC-GG-3'; 3'-primer, 5'-AGT-GAA-CAT-TAC-AGA-TTT-ATC-CC-3'; 437 bp), MIP1-
(5'-primer, 5'-GTC-TTC-TCA-GCG-CCA-TAT-GG-3'; 3'-primer, 5'-GGC-ATT-CAG-TTC-CAG-GTC-AG-3'; 225 bp), and KC (5'-primer, 5'-TTT-GGA-CAA-TTT-TCT-GAA-CC-3'; 3'-primer, 5'-GAT-TCA-CCT-CAA-GAA-CAT-CC-3'; 123 bp).
cDNAs were amplified by repeated cycles at 95°C for 20 s, at the appropriate annealing temperature for 45 s and at 75°C for 45 s. PCR amplifications were performed during the exponential phase of amplification, with a nonsaturating number of cycles (15 to 35) being used for each cytokine. Amplification of ß2-microglobulin (17 cycles) was used as a control. PCR products were run on 1.2% agarose gels supplemented with ethidium bromide, and their sizes were evaluated using molecular size standards (123-bp ladder [Life Technologies]). For each cytokine, the relative expression of the mRNA was determined in infected and control BALB/c and C.CB mice by a semiquantitative analysis. PCR products from a given set of samples were analyzed for the cytokine and ß2-microglobulin genes. The amounts of PCR products obtained in exponential phases of amplification were quantified by measuring the intensity of fluorescence of amplicons by densitometric analysis (Kodak Image station 440 [NEN Life Science Products, Paris, France]). The ratio of the cytokine amplification product to the ß2-microglobulin amplification product, which is representative of the relative expression of the cytokine mRNA in the sample, was then calculated (28). For a given cytokine, the samples were compared, and gene expression was considered to be significantly different if the calculated values varied by a factor of at least 1.5.
Statistics. Analysis of variance (two-factor ANOVA) was performed on data followed by Student-Newman-Keuls or Scheffe tests for orthogonal analysis. These tests were done using Superanova (Abacus Concepts, Berkeley, Calif.).
|
|
|---|
![]() View larger version (13K): [in a new window] |
FIG. 1. Growth of B. melitensis in spleens and livers of BALB/c (Nramp1s) ( ) and congenic C.CB (Nramp1r) () mice. Mice (n = 8 mice per time point) were inoculated i.p. with 104 CFU of Brucella. Results are expressed as means ± SD. The number of Brucella organisms per spleen was significantly larger in C.CB mice than in BALB/c mice during the first week p.i. (P < 0.005).
|
![]() View larger version (26K): [in a new window] |
FIG. 2. Isolation of B. melitensis associated with splenocytes from infected BALB/c and C.CB mice. Splenocytes were isolated on day 5 or 14 after i.p. inoculation of mice with B. melitensis. The number of Brucella organisms per spleen was counted before (open bars) and after (shaded bars) in vitro incubation with gentamicin. Results are expressed as means ± SD (n = 4 mice per time point). *, statistically different from BALB/c mice (P < 0.01); NS, not significant
|
Effect of Nramp1 expression on the inflammatory response induced by B. melitensis infection. In the absence of any direct effect of Nramp1 expression on the intracellular replication of B. melitensis in isolated macrophages infected in vivo or in vitro, we analyzed the pleiotropic effect of Nramp1 on the inflammatory response in the spleen. Indeed, spleen weight increased steadily after infection with B. melitensis (on average from 100 to 600 mg) in both mouse strains. However, between 14 and 28 days p.i., splenomegaly was significantly greater in C.CB mice than in BALB/c mice (P < 0.01) (Fig. 3). An increase in liver weight was also noticed and was associated with significant greater hepatomegaly in the C.CB mice (P < 0.05).
![]() View larger version (14K): [in a new window] |
FIG. 3. Kinetics of spleen and liver weights after B. melitensis infection of BALB/c ( ) and C.CB () mice. Results are expressed as means ± SD (n = 8 mice per time point). The increase in spleen and liver weights in the C.CB mice was significantly greater than that in the BALB/c mice (P < 0.01 and P < 0.05 for spleen and liver, respectively).
|
, IL-12p40, IFN-
, iNOS, and IL-10 mRNAs were expressed in the splenocytes of infected mice but not in those of the control mice. In contrast, no significant induction of expression of MIP-1
, IL-18, IL-2, and IL-4 transcripts (ratio of amplification, < 1.5) could be demonstrated, with similar levels of mRNAs being observed in BALB/c and C.CB mice regardless of whether they were infected (Fig. 4A). A semiquantitative analysis showed that the mRNA expression of TNF-
, IL-12p40, iNOS, and IL-10 was not significantly different in the two strains of mice but that the level of IFN-
mRNAs was two to three times higher in infected C.CB mice than in infected BALB/c mice when their relative expression of mRNAs was compared. In contrast, the transcript for the chemokine KC was specifically amplified in the infected C.CB mice alone. A similar pattern of cytokine mRNAs was observed when spleens were analyzed 14 days p.i., except for TNF-
mRNAs, which no longer displayed induction. Nevertheless, the relative expression of mRNAs for each cytokine, representative of four mice for each of the different groups, showed that the mRNAs were more strongly expressed in infected C.CB mice than in infected BALB/c mice: KC (sevenfold increase), IL-12p40 (fivefold increase), and IFN-
(two- to threefold increase) (Fig. 4B).
![]() View larger version (39K): [in a new window] |
FIG. 4. Cytokine mRNAs expression in spleens of B. melitensis-infected or uninfected BALB/c and C.CB mice. (A) The gene expression of different cytokines was analyzed by RT-PCR performed on mRNAs of infected BALB/c (lanes 3) and C.CB (lanes 4) mice and their respective controls (lanes 1 and 2). (B) Four mice per group were sacrificed at 5 and 14 days p.i., and spleen cells were isolated and pooled for mRNA preparation. For each cytokine, the relative expression of mRNAs was calculated from the relative intensity of the amplification product, determined by densitometric analysis, with respect to that of the housekeeping gene, ß2-microglobulin. Values for the expression of KC, IL-12, IFN- , and iNOS mRNAs 14 days p.i. are shown.
|
![]() View larger version (35K): [in a new window] |
FIG. 5. Fluorescence-activated cell sorter analysis of RB6-8C5 and BM8 marker expression in splenocytes from BALB/c and C.CB mice after infection by B. melitensis. Control and infected mice were sacrificed 5 or 14 days p.i. Spleen cells were labeled with MAbs and fluorescence-conjugated secondary antibody, and positive cells were analyzed in total gated leukocytes. Results are expressed as means ± SD of the number of positive cells per spleen (n = 4). Differences in the numbers of RB6-8C5+ and BM8+ cells were not statistically significant for infected BALB/c and C.CB mice.
|
|
|
|---|
Brucella organisms survive and replicate in host macrophages, where the Nramp1 gene is expressed. Experiments were therefore performed to investigate the relationship between the intracellular localization of Brucella and the different patterns of infection observed in BALB/c and C.CB mice during the early phases of infection. The number of Brucella organisms associated with cells isolated from infected spleens was still larger in C.CB than in BALB/c mice after treatment with gentamicin, an antibiotic able to kill extracellular Brucella. These findings confirmed data obtained using homogenized spleens and demonstrated that the early effect on infection triggered by the Nramp1 gene occurred in intracellular Brucella organisms.
The in vitro replication of Brucella was similar in macrophages isolated from the two strains of mice infected in vivo and in Nramp1 RAW264.7 transfectants. This excludes any direct involvement of the Nramp1 gene in the intracellular replication of Brucella in macrophages. Such a discrepancy between in vivo and in vitro data has been reported for C57BL/10 and BALB/c mice, strains which are resistant and susceptible, respectively, to B. abortus infection. The macrophages of these mice show no difference in their ability to control Brucella replication (39). In contrast to Nramp1 RAW264.7 transfectants, which inhibit the intracellular replication of S. enterica serovar Typhimurium (20), stable transgenes of the resistance-associated allele of the bovine NRAMP1 gene in RAW264.7 cells control the in vitro replication of B. abortus but not of S. enterica serovar Dublin (3). These apparently conflicting data suggest that (i) the polymorphism of the resistance-associated allele of bovine NRAMP1 gene affects the process of intracellular replication of Brucella but not of Salmonella; (ii) the point mutation in the coding region of the murine Nramp1 gene, resulting in a single amino acid substitution of Gly by Asp at position 169, does not affect the in vitro intracellular replication of B. melitensis in macrophages; and (iii) genes other than Nramp1 are involved in the natural resistance of mice to Brucella. Nevertheless, transgenic mice could provide a useful model for studying the specific effects of the expression of the resistance-associated allele of the bovine NRAMP1 gene on the outcome of Brucella infection.
Several studies have reported a link between the antibacterial activity of the murine Nramp1 gene and the phagolysosomal pathway and, more precisely, with phagosomal acidification (4, 22). The ability of Brucella to survive in acidic phagosomes and the requirement for early acidification of this compartment for their survival within macrophages (35) could explain why the expression of the Nramp1 gene does not affect the intramacrophage development of Brucella. One explanation could be that Brucella exploits the ion transport function of Nramp1 to its own advantage, as was suggested recently for F. tularensis (26). Indeed, numerous genera of bacteria express genes encoding the bacterial ortholog of the eukaryotic Nramp protein, known as MntH (from the proton-dependent manganese transporter) (12). A homolog of the MntH gene of Agrobacterium tumefaciens is present in the genome of B. melitensis (13) and displays 68% nucleotide identity (GenBank accession number AE009498). This suggests that a divalent cation transporter homolog of the Nramp1 family could be expressed in Brucella and that this could compete for metal ions within the intraphagosomal vesicles of macrophages.
In the absence of any direct effect of Nramp1 expression on the in vitro intramacrophage replication of B. melitensis, we analyzed in vivo data by considering the pleiotropic effects of the Nramp1 gene on the inflammatory response and macrophage activation pathways (5). The infection of BALB/c and C.CB mice resulted in an inflammatory response characterized by splenomegaly, which began 5 days p.i., and in the influx of neutrophils (RB6-8C5+ cells) and macrophages (BM8+ cells) into the spleen. This response was amplified in C.CB mice, in line with the higher bacterial load. In both strains of mice, in accordance with the onset of splenomegaly observed on day 5 p.i., the induction of TNF-
, iNOS, IL-12p40, and IFN-
, but not of IL-2 and IL-4 mRNAs, was typical of the inflammatory and Th1 cytokine responses required to impair Brucella infection (30). The induction of IL-10 mRNAs was probably a consequence of the inflammatory response, since IL-10 is required to control this response. This pattern of cytokine expression was maintained during the second week of infection, except for TNF-
, which was no longer expressed. IFN-
and IL-12p40 mRNAs were more intensively induced in the infected C.CB mice than in the infected BALB/c mice. Despite the early enhanced susceptibility to B. melitensis infection of the C.CB mice, they develop an amplified Th1 cytokine response, which fits in with the greater capacity of macrophages carrying the Nramp1r allele to promote the Th1 cytokine response in response to Leishmania infection than of macrophages carrying the Nramp1s allele (41). In addition to Th1 cytokines, a dramatic difference was observed in the expression of the C-X-C chemokine KC, which was induced (on day 5 p.i.) in the C.CB mice only. This differential induction of the KC chemokine in infected Nramp1r mice is not surprising, since its expression is known to be controlled by the Nramp1 gene (38). KC is a potent inducer of neutrophil migration (9); therefore, its production could potentiate the recruitment of neutrophils, as observed in infected C.CB mice. This recruitment could be amplified by neutrophils themselves since high-level expression of Nramp1 has been recently reported for these cells (10). The polymorphism of other genes beside the Nramp1 gene, transferred in congenic C.CB mice, like IL-8RB (http://www.informatics.jax.org), could also participate in the recruitment of neutrophils. We detected a higher recruitment of neutrophils in the spleens of infected C.CB mice than in those of BALB/c mice. Therefore, the temporary susceptibility of C.CB mice to B. melitensis infection could be related to a greater recruitment of phagocytic cells. Since no direct effect of Nramp1 on the intramacrophage replication of Brucella organisms was observed, this suggests that the increased inflammatory response in spleens of C.CB mice did not correlate with the increased control of splenic B. melitensis. However, the increased inflammatory response in spleens of C.CB mice, maintained for 3 weeks p.i., could also be a response to the increased bacterial load in these spleens; phagocytic cells represent both host and effector cells in Brucella infection. In any case, the subsequent development of a pronounced compensatory Th1 response by C.CB mice counteracts and stabilizes the infection in the spleen. The inflammatory response and Th1 response stabilize the number of Brucella organisms in the spleen at a similar level in the two strains of mice, but in both cases they do not lead to the clearance of the infection.
In conclusion, Nramp1 gene expression confers only a temporary advantage for B. melitensis survival in mice and is of limited importance for the natural resistance of mice to Brucella.
|
|
|---|
gene in early response to Salmonella infection. Microbes Infect. 3:1-6.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»