Innate Immunity to Intraphagosomal Pathogens Is Mediated by Interferon Regulatory Factor 8 (IRF-8) That Stimulates the Expression of Macrophage-specific Nramp1 through Antagonizing Repression by c-Myc*

Macrophages are a central arm of innate immune defense against intracellular pathogens. They internalize microbes into phagosomes where the invaders are being killed by oxygen and nitrogen reactive species. Despite this battery of antimicrobial molecules, some are able to thrive within the phagosome thus termed intraphagosomal pathogens among which are Salmonella, Leishmania, and Mycobacteria. In mice, a single dominant gene termed Nramp1/Slc11a1 controls innate resistance to such pathogens. This gene is expressed exclusively in myeloid cells. Previously, we have shown that the restricted expression of Nramp1 is regulated by a myeloid cell-specific transcription factor termed IRF-8/ICSBP. It is demonstrated here that the induction of Nramp1 expression in activated macrophages is accompanied by a promoter shift from a repression state elicited by c-Myc to an activation state elicited by the induction of IRF-8 in activated macrophages. This transition from repression to activation is facilitated by a competitive protein-protein interaction with the transcription factor Miz-1. To show that IRF-8 is directly involved in the elimination of intraphagosomal pathogens through the regulation of Nramp1 gene expression, we bred wild type as well as IRF-8 and Nramp1 null mouse strains and examined macrophages derived from bone marrow and peritoneum. Our results clearly show that the absence of IRF-8 and Nramp1 leads to the same phenotype; defective killing of intraphagosomal Salmonella enterica serovar typhimurium and Mycobacterium bovis. Thus, interplay between repression and activation state of the Nramp1 promoter mediated by IRF-8 provides the molecular basis by which macrophages resist intraphagosomal pathogens at early stage after infection.

Macrophages are a central arm of innate immune defense against intracellular pathogens. They internalize microbes into phagosomes where the invaders are being killed by oxygen and nitrogen reactive species. Despite this battery of antimicrobial molecules, some are able to thrive within the phagosome thus termed intraphagosomal pathogens among which are Salmonella, Leishmania, and Mycobacteria. In mice, a single dominant gene termed Nramp1/Slc11a1 controls innate resistance to such pathogens. This gene is expressed exclusively in myeloid cells. Previously, we have shown that the restricted expression of Nramp1 is regulated by a myeloid cell-specific transcription factor termed IRF-8/ICSBP. It is demonstrated here that the induction of Nramp1 expression in activated macrophages is accompanied by a promoter shift from a repression state elicited by c-Myc to an activation state elicited by the induction of IRF-8 in activated macrophages. This transition from repression to activation is facilitated by a competitive protein-protein interaction with the transcription factor Miz-1. To show that IRF-8 is directly involved in the elimination of intraphagosomal pathogens through the regulation of Nramp1 gene expression, we bred wild type as well as IRF-8 and Nramp1 null mouse strains and examined macrophages derived from bone marrow and peritoneum. Our results clearly show that the absence of IRF-8 and Nramp1 leads to the same phenotype; defective killing of intraphagosomal Salmonella enterica serovar typhimurium and Mycobacterium bovis. Thus, interplay between repression and activation state of the Nramp1 promoter mediated by IRF-8 provides the molecular basis by which macrophages resist intraphagosomal pathogens at early stage after infection.
Mammals host defense against pathogens depends on both innate and adaptive immunity. Macrophages are essential com-ponents of innate immunity and play a key role in host defense mechanisms against invading pathogens. These invading foreign bodies activate the macrophages, which in response engulf and subsequently entrap the pathogens in the phagosome/lysosome compartment. In this specialized compartment, the pathogens are subjected to massive attack by reactive oxygen and nitrogen intermediates (1). In addition, macrophages mediate innate resistance to host infection by several antigenically unrelated intraphagosomal pathogens including Salmonella, Leishmania, and Mycobacterium. This is controlled by a single dominant gene termed solute carrier family 11 member a1 (Slc11a1) 2 also known as natural resistance associated macrophage protein 1 (Nramp1), or Ity/Lsh/Bcg. Nramp1 is a proton/divalent cation antiporter with a unique role in innate resistance to intraphagosomal pathogens and autoimmune disease (for review, see Refs. [2][3][4]. The exact function of Nramp1 as iron and divalent cation transporter is still controversial. It either functions to increase transphagosomal Fe 2ϩ catalyzing the Haber-Weiss/Fenton reaction to generate the highly toxic hydroxyl radical essential for macrophage bactericidal activity. On the other hand, it is thought to deprive the intraphagosomal bacterium of Fe 2ϩ and other divalent cations, which are critical for its survival in the phagosome (2). In mouse, Nramp1 has two alleles; one allele, Gly-169, restricts pathogen growth, whereas the second allele, Asp-169 (functionally null), is permissive for pathogen growth (5). Thus, a single gene product, Nramp1, is indispensable for innate immunity to intraphagosomal pathogens. Importantly, Nramp1 is exclusively expressed in monocyte/macrophage cells (5), whereas its family member Nramp2 is ubiquitously expressed. The transcription factor termed Myc interacting zinc finger protein 1 (Miz-1) mediates the activation of Nramp1 that can be repressed by c-Myc (25). However, this interaction between Miz-1 and c-Myc does not explain the macrophage-restricted expression of Nramp1 and its inducibility by interferon-␥ (IFN-␥) and LPS in comparison to its ubiquitously expressed homologue Nramp2.
Recently, we have demonstrated that Miz-1 interacts with another transcription factor termed interferon regulatory factor-8 (IRF-8) previously known as ICSBP. Furthermore, we have shown that these two factors in conjunction with a third partner, PU.1, synergistically activate Nramp1 (6). IRF-8 is expressed mainly in immune cells and its expression can be further induced by IFN-␥ and Toll receptor ligands such as LPS or CpG (7). This factor is a key element for the differentiation of myeloid progenitor cells toward macrophages and for mature macrophage activity. Accordingly, IRF-8 Ϫ/Ϫ mice exhibit clinical manifestation that resembles the human chronic myelogenous leukemia. Thus, IRF-8 drives bipotential myeloid progenitor cells toward mature macrophages while inhibiting the differentiation pathway toward granulocytes (8). In addition, it was shown that IRF-8 is an essential factor for proper function of mature macrophages. For example, IRF-8 null mice fail to mount Th1-mediated immune response due to lack of expression of IL-12 p40 subunit (9). IRF-8 belongs to a family of 9 cellular members sharing significant similarity at the DNA binding domain. Unlike other IRF members, IRF-8 is capable of binding to target DNA sequence only following association with other IRF or non-IRF transcription factors via the IRF association domain (10). IRF-8 associates with transcription factors like IRF-1 and IRF-2 as well as non-IRF members such as PU.1. The latter belongs to the Ets family of transcription factors and is indispensable for the development of lymphoid and myeloid cells (11). The interacting partner with IRF-8 dictates not only the DNA binding site but also the transcriptional activity, e.g. activation or repression. Through the formation of such heterocomplexes IRF-8 has an important role in the expression of macrophage genes such as the phagosomal components, phagocyte oxidase complex (phox) (12), and iNOS (13), or the proinflammatory cytokines IL-12 (9), IL-23 (14), IL-18 (15), and IL-1␤ (16).
Previously, we have shown that IRF-8 can interact with Miz-1 only in immune cells (6). This interaction leads to synergistic activation of the Nramp1 promoter that is not observed in the non-immune cell line. This synergistic interaction is further enhanced with PU.1. These results clarified the molecular basis for the macrophage-restricted expression of Nramp1 and its inducibility by IFN-␥ and LPS. In this article, we wished to further explore the Nramp1 promoter with respect to the identification of the exact binding sites for IRF-8 and PU.1. In addition, the competitive interplay between c-Myc and IRF-8 for the association with Miz-1 and the effect on the transcriptional activity of the Nramp1 promoter was investigated. Finally, a direct regulatory cascade linking IRF-8 and Nramp1 sequential expression to macrophages susceptibility/ resistance to intraphagosomal pathogens such as Salmonella enterica serovar typhimurium and Mycobacterium bovis (bacille Calmette-Guerin (BCG)) is described.

EXPERIMENTAL PROCEDURES
Animals-The mouse strains, C57BL/6J, 129/Sv (both were from Harlan Biotech, Israel), and C57BL/6J IRF-8 Ϫ/Ϫ (17) were maintained in individually ventilated cages under pathogenfree conditions. The C57BL/6J IRF-8 Ϫ/Ϫ mice were backcrossed with 129/Sv for 8 generations to replace the defective Nramp1 allele generating a new knock-out mouse strain 129/Sv IRF-8 Ϫ/Ϫ . The presence of defective IRF-8 and functional Nramp1 were validated by PCR of genomic DNA from mouse tails. All animal work conformed to the guidelines of the animal care and use committee of the Technion.
Cell Lines-Murine bone marrow-derived RAW264.7 macrophage cell line was obtained from ATCC (Manassas, VA), CL2 macrophage cells, established from IRF-8 knock-out mice, were previously reported (9), R37 and R21 cells (constitutively expressing functional Nramp1 or antisense to Nramp1, respectively) were a kind gift from Dr. Barton (Southampton University, United Kingdom) (18). Cells were maintained in RPMI 1640, supplemented with 40 M ␤-mercaptoethanol, 10% fetal calf serum, and antibiotics. 5 ng/ml M-CSF and CSF (R&D Systems, Minneapolis, MN) were added to the growth medium of CL2 cells and 50 g/ml of gentamycin was added to the growth medium of R37 and R21 cells. Treatment with IFN-␥ and LPS and killing assays were performed as described for peritoneal macrophages. NIH3T3 cells (ATCC) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat inactivated fetal calf serum and antibiotics.
Isolation of Bone Marrow and Peritoneal Macrophages-Bone marrow cells were isolated from femurs of the various mouse strains and cultured for 7 days in Dulbecco's modified Eagle's medium supplemented with 30% CCL1 cell culture supernatant, 20% heat inactivated fetal calf serum, and antibiotics in nontissue culture plates. Peritoneal macrophages were harvested as described previously (6). 2.5 ϫ 10 7 cells were plated in tissue culture Petri dishes (100 mm) and kept at 37°C and 6% CO 2 for 4 h and then non-adherent cells were removed by 2 washes with PBS.
Intracellular Pathogens-S. enterica serovar typhimurium 14028 (ATCC) stably expressing green fluorescent protein (GFP) was kindly obtained from Dr. Yaron (Technion, Israel). Following 18 h growth (LB medium without antibiotics) the bacterial culture was diluted 1:10 and A 600 was measured. Salmonella concentration was calculated as colony forming unit (CFU) per ml using 1 OD ϭ 1.2 ϫ 10 9 CFU/ml. The cells were washed once with PBS and resuspended in mammalian culture medium. Macrophages were infected at multiplicity of infection (m.o.i.) of 15-30. The Leishmania donovani stably expressing eGFP was a kind gift of Dr. Zilberstein (Technion, Israel). The cells were maintained in M199 culture medium with 10% heat-inactivated fetal calf serum and antibiotics at 26°C that favors flagellate promastigote form as described (19). For the assays, the parasites were collected from end-log-phase cultures, centrifuged at 800 ϫ g, washed with PBS, counted after fixation with paraformaldehyde, and adjusted to the required parasite m.o.i. of 10.
Macrophage Killing Assay-Killing assay with S. enterica serovar typhimurium was performed as described by Monak et al. (20) with minor modifications. Macrophages (either bone marrow derived or peritoneal) were washed twice with ice-cold PBS and the cells were gently scraped and plated at 3 ϫ 10 5 cells/well in 24-well plates in a medium without antibiotics.
The cells were either not treated or treated for 18 h with 100 units/ml IFN-␥ (CytoLab, Rehovot, Israel) and 50 ng/ml LPS (Sigma). Cells were infected in triplicates with S. enterica serovar typhimurium 14028 expressing at m.o.i. of 15-30. The 24-well plates were centrifuged at 250 ϫ g for 5 min and placed in 5% CO 2 incubator at 37°C for 40 min. This point in the experiment was determined as the zero point. At this stage gentamycin (50 g/ml) was added to kill non-phagocytized bacteria and to assess intracellular growth. At the indicated time intervals post-infection, cells were washed 3 times with PBS, lysed with 200 l of 1% deoxycholate, serially diluted, plated onto LB agar plates, and CFU (viable bacteria) were determined.
For killing assay with M. bovis (BCG), bone marrow-derived macrophages from various mouse strains were plated and activated with IFN-␥ and LPS as described above. Macrophages were then infected with live M. bovis (BCG) at m.o.i. of 10 for 4 h, the adherent cells were washed twice, and fresh growth medium was added. Cells were harvested and lysed in 2 ml of sterile water, and 100-l aliquots were used for preparing serial dilution of the bacteria. The diluted lysates were placed on oleic acid-albumin-dextrose-catalase-supplemented Middlebrook 7H11 Bacto-agar (Difco), and colonies were counted visually after 14 days.
Plasmids-Mammalian expression vectors encoding for Miz-1, IRF-8, IRF-1, and PU.1 were all described previously (6). The expression plasmid for c-Myc in pCDNA3.1 was a gift of Dr. Eilers (Marburg University, Germany). The reporter construct Ϫ1555 containing murine Nramp1 prompter in the reporter vector pGL3 was a gift from Dr. Barton (Southampton University). The various 5Ј progressive Nramp1 promoter deletion constructs were generated by PCR and following sequence validation were cloned into pGL3 basic (Promega). Primer sequence is available upon request. Site-directed mutagenesis to the IFN-stimulated response element (ISRE, the binding site for IRF-8) and the PU.1 binding sites within the mouse Nramp1 promoter was performed using the QuikChange site-directed mutagenesis kit (Stratagene). Primer sequence is available upon request.
Cells Transfections and Reporter Gene Assays-Assays were performed exactly as described before (6) using the Dual Luciferase assay kit (Promega). Luciferase activities were determined and normalized for transfection efficiency. -Fold of synergism was calculated as the ratio between the -fold of activation elicited by the two transfected factors divided by the sum of the -fold of activation for each factor alone. Each experiment was repeated at least 3 times.
Bifluorescence Complementation (BiFC)-The plasmid IRF-8-YCC was previously described (23). The plasmid Miz-1-YNC was generated by PCR amplification of the coding region of Miz-1 with primers adding EcoRI and HindIII at the 5Ј and 3Ј ends, respectively, and cloned into the corresponding restriction sites in the plasmid YNC (23). 3.5 ϫ 10 4 NIH3T3 cells/ chamber were seeded in sterile 4-well slide chamber. 18 h later, cells were transfected by Lipofectamine (Invitrogen) with 0.25 g of the YFP constructs and 0.5 g of pHcRed1-C1 (Clontech) for 24 h. The later served as a control reporter plasmid to confirm that BiFC took place within a nucleus of a transfected cell.
In addition, 0.75 g of plasmid encoding for c-Myc was cotransfected as indicated. The cells were allowed to stand at 30°C for 1 h, and YFP and red fluorescent protein signals were viewed under confocal microscopy.
Retroviral Transduction-pMSCV-IRES-EGFP is a retroviral expression vector that has a bicistronic cassette that enables expression of IRES-driven EGFP, with no encoded gene in the first position. It was generated by digesting the bicistronic cassette from pIRES2-EGFP (Clontech) with XhoI and MfeI and cloned into pMSCV digested with XhoI and EcoRI. To generate the retroviral pMSCV-IRF8-IRES-EGFP expression vector, the cDNA for human IRF-8 was digested with PmeI and MfeI and cloned into the HpaI and EcoRI in the retroviral vector pMSCV-IRES-EGFP. Retroviruses were generated as described (25) by seeding 4 ϫ 10 6 293FT cells in a 10-cm dish. The next day, cells were transfected using DNA-calcium phosphate precipitate containing 2.5 g of plasmid pMD.G encoding vesicular stomatitis virus G protein, 7.5 g of plasmid pMD.OGP encoding gag-pol, and 10 g of the retroviral expression constructs. 48 h after transfection, the viral supernatant was collected, centrifuged at 800 ϫ g, and used to infect bone marrow macrophages that were seeded at 7.5 ϫ 10 6 cells/well in a 12-well plate in a final volume of 1.5 ml. 24 h later, 0.5 ml of medium was remove and replaced with 0.5 ml of freshly prepared viruses and Polybrene was added to a final concentration of 8 g/ml. The plates were spinoculated at 1200 ϫ g for 90 min at 33°C. After 24 h, 1 ml of the medium was replaced with fresh medium and 1 day later transfection efficiency was monitored under a fluorescent microscope and the transduced cells were selected with puromycin (3 g/ml) for an additional 7 days and then used for Salmonella killing assay.
Statistical Analysis-Each experiment was repeated three times yielding similar results and statistical significance was determined by Student's t test with a p Ͻ 0.05.

RESULTS
Deletion Analysis of the Nramp1 Promoter Revealed the Binding Sites for IRF-8 and PU.1-Our previous studies provided the molecular basis for the restricted expression of Nramp1 in activated macrophages. This is due to synergistic interaction between Miz-1, IRF-8, and PU.1. However, the exact binding sites for the myeloid-specific factors were not determined. Computer aided analysis identified several putative binding sites for IRF-8 (termed ISRE) and PU.1 along the Ϫ1555 bp of the Nramp1 promoter region. To test which of the putative binding sites serve as ISRE and PU.1 binding sites, reporter gene constructs harboring progressive 5Ј promoter deletions were generated (see illustrations in Fig. 1A). RAW264.7 cells were cotransfected with these reporter constructs together with Miz-1 and IRF-8 (Fig. 1B, front row) as well as with PU.1 (Fig. 1B, back row). As expected, the full-length promoter construct (Ϫ1555) was synergistically activated by Miz-1 and IRF-8 (Fig. 1B, black bar, front row). This activation was further enhanced by PU.1 (Fig. 1B, black bar, back row). The synergistic effect of Miz-1 and IRF-8 was also noted with the reporter constructs Ϫ167 and Ϫ139 (gray bar and striped bar, front row, respectively). However, removing just four bases from Ϫ139 to Ϫ135 abolished PU.1 enhancement (dotted bar, back row). This 4-bp deletion, TTCC/GGAA, harbors a consensus binding site for PU.1. As expected, transfection with the Ϫ96 reporter construct exhibited no effect with PU.1 (checkered bar, back row) and in addition the synergistic effect of both IRF-8 and Miz-1 was abolished (checkered bar, front row). In this reporter construct, half of the palindrome GTTTCGAAAC was deleted. This palindrome fits the consensus of an ISRE and suggests that it serves as an IRF-8 binding site.
Site-directed mutagenesis was performed for these two ISRE and PU.1 binding sites. As seen in Fig. 1C, mutating the PU.1 binding site did not affect Miz-1 and IRF-8 synergistic activation of this mutated construct but the additional enhancement with PU.1 was abolished. Similarly, mutating the ISRE binding site resulted in loss of the synergistic effect between IRF-8 with Miz-1. These results allowed us to assign these two binding sites as PU.1 and ISRE, respectively.
To establish our findings that IRF-8 regulates Nramp1 expression by binding to its promoter, RAW264.7 cells were either not activated or activated with IFN-␥ for 12 h and ChIP assays were preformed as described under "Experimental Procedures." The primers used for the PCR were designed to include the ISRE and PU.1 binding sites on the Nramp1 pro- 24 h later cells were harvested, and relative luciferase activities were determined and the -fold of synergism was calculated as described under "Experimental Procedures." C, site-directed mutagenesis was performed to the ISRE (adjacent to position Ϫ96) and the PU.1 binding site (near position Ϫ139) on the Ϫ139 Nramp1 reporter construct generating mutated reporter constructs termed Ϫ139ISREmut and Ϫ139PU.1mut, respectively (see Fig. 3A for detailed illustration). RAW264.7 cells were cotransfected with Miz-1, IRF-8, and PU.1 as described under panel A and indicated in the figure. Relative luciferase activities were determined and the -fold of synergism was calculated as described under "Experimental Procedures." D, semiquantitative ChIP was performed in RAW264.7 cells before and following 12 h of activation with IFN-␥ and LPS using the indicated antibodies. E, quantitative ChIP to determine IRF-8 binding activity to Nramp1 promoter was performed as described under "Experimental Procedures." moter. The binding of IRF-8 to Nramp1 promoter was evident by semi-quantitative PCR only under IFN-␥ induced conditions (Fig. 1D). The relative binding of IRF-8 to the Nramp1 promoter in the induced cells was increased by almost 5-fold (Fig.  1E). The specificity of the ChIP assay was evaluated using goat immunoglobulins (IgG, negative control) and acetyl-specific histone 3 antibody (positive control, data not shown). These results support the in vitro findings and provide in vivo evidence for the recruitment of IRF-8 to the Nramp1 promoter in the IFN-␥-activated macrophage cell line.
c-Myc Negates IRF-8 and Miz-1 Transcriptional Synergy on Nramp1 Promoter-As shown above, Nramp1 is induced in activated macrophages through synergistic association between IRF-8 and Miz-1. On the other hand, other reports demonstrated that c-Myc represses Nramp1 transcription by binding to Miz-1 and displacing the p300 co-activator, a paradigm for the repressive action of c-Myc (26 -28). This suggested that Miz-1 interaction with either c-Myc or IRF-8 might affect Nramp1 enhanceosome activity leading to repression or activation, respectively.
To test the effect of c-Myc on the synergistic activation of the Nramp1 promoter by IRF-8 and Miz-1, reporter gene assays were performed. The Nramp1 reporter construct (Ϫ1555) was co-transfected with expression vectors for IRF-8, Miz-1, and c-Myc. As previously shown (6), transfection of Miz-1 alone was sufficient to significantly activate the promoter (Fig. 2A,  lane 2), whereas IRF-8 and c-Myc alone exhibited no effect ( Fig.  2A, lanes 3 and 4, respectively). As expected, cotransfection of Miz-1 and IRF-8 resulted in a significant activation of the promoter (ϳ12-fold, Fig. 2A, lane 5). Interestingly, cotransfection of increasing amounts of c-Myc reporter construct inhibited Miz-1 and IRF-8 synergistic activation reaching the reporter level similar to that of Miz-1 and c-Myc alone ( Fig. 2A, compare lanes 6 -10 with lane 11, respectively). In analogy, the repression effect of c-Myc on Miz-1 (Fig. 2, lanes 11 and 2, respectively) could be partially alleviated by cotransfection of increasing amounts of IRF-8 expression vector ( Fig. 2A, lanes 12-14). These results suggest that c-Myc and IRF-8 compete over the interaction with Miz-1. Whereas the first exert repression the latter negates this activity. Furthermore, it suggests that the proteinprotein interaction between Miz-1 and c-Myc is stronger than that of Miz-1 and IRF-8 because IRF-8 could not completely reverse the repression effect of c-Myc.
BiFC assay was employed to visualize Miz-1, IRF-8, and c-Myc interactions in living cells (29). Expression vectors for IRF-8 fused to the C-terminal region of the YFP (IRF-8-YCC) and for Miz-1 fused to the N-terminal region of the YFP (Miz-1-YNC) were constructed. These plasmids were transiently transfected into NIH3T3 cells either together or alone. As seen in Fig. 2B, expression of both IRF-8-YCC and Miz-1-YNC resulted in nuclear fluorescence of the transfected cells (ϳ10% efficiency) demonstrating their interaction in living NIH3T3 cells. This supports our previous demonstrations of such interaction using mammalian two-hybrid and co-immunoprecipitation assays (6). Cotransfection of the expression vector encoding for c-Myc resulted in a sharp decrease of fluorescent nuclei, indicating that c-Myc interfered with IRF-8 and Miz-1 interaction (Fig. 2B). Transfection of expression vectors for IRF-8-YCC or Miz-1-YNC alone did not lead to any fluorescence (data not shown). Transfection efficiency was monitored by cotransfection of a red fluorescent protein expression vector (for details see "Experimental Procedures"). Together, these results suggest that interplay between c-Myc and IRF-8 expression following IFN-␥ and LPS stimulation modulate the transcriptional activity of Nramp1 in macrophages.

IRF-8 and PU.1 Are Important Regulators of Nramp1 Activation during Infection with Intraphagosomal Pathogens-We
next wanted to study the significance of IRF-8 and PU.1 on Nramp1 expression during infection with intraphagosomal pathogens. For that purpose, RAW264.7 cells were transfected with the Nramp1 reporter construct (Ϫ139) or with the same construct harboring point mutations in the ISRE or PU.1 binding sites (see schematic illustration in Fig. 3A). Twenty h following transfection, cells were either not infected or infected with S. enterica serovar typhimurium (expressing GFP) at m.o.i. of 15 for 5 h. Infection efficiency was evaluated using fluorescent microscope, whereas the activation of the different Nramp1 promoters was measured by luciferase assay. As shown in Fig. 3B, the wild-type (WT) Nramp1 promoter is activated during S. enterica serovar typhimurium infection (nor- malized to uninfected cells) by an average of 1.9-fold (black bar). However, the mutated promoters remained at their basal activity level after infection and were not induced by the S. enterica serovar typhimurium (white and gray bars). These results show that intact ISRE and PU.1 binding sites are essential for Nramp1 activation and strongly suggests that IRF-8 and PU.1 are essential factors for Nramp1 expression in macrophages during S. enterica serovar typhimurium infection.
To further confirm the importance of IRF-8 for Nramp1 promoter activation during intraphagosomal infection, we used the protozoa L. donovani in similar reporter gene assays. RAW264.7 cells (WT for IRF-8 expression) and CL-2 cells (derived from IRF-8 knock-out mice (9)) were transfected with the Nramp1 reporter construct (Ϫ1555) and 4 h later the cells were infected with L. donovani promastigotes (expressing GFP) at m.o.i. of 10 for an additional 20 h. Infection efficiency was evaluated using a fluorescent microscope. Similar infection efficiencies were observed for both cell lines (Fig. 2C). Promoter activity was measured by luciferase assay and results were normalized to uninfected cells. Similar to the results observed with S. enterica serovar typhimurium, L. donovani activated Nramp1 in an IRF-8-dependent manner. Significant activation following infection was observed only in RAW264.7 cells (ϳ1.9-fold, Fig.  3B, black bars) and not in CL-2 cells, which are ablated for IRF-8 (Fig. 3B, white bars).
Nramp1 Is Essential for the Elimination of S. enterica Serovar typhimurium in Infected RAW264.7 Cells-Subsequent to establishing the molecular basis for induced expression of Nramp1 by IRF-8 and PU.1 in infected cells, we wanted to study its importance in macrophage-mediated killing of pathogens at early stages of infection. For that purpose, S. enterica serovar typhimurium killing assay was performed in vitro with RAW264.7 cells (expressing a mutated allele of Nramp1 (D169)), which stably express either functional Nramp1 (R37 cells), or Nramp1-antisense (R21 cells). These two cell lines were activated with IFN-␥ and LPS for 20 h and then infected with S. enterica serovar typhimurium expressing eGFP as described under "Experimental Procedures." The infection efficiency was evaluated under fluorescent microscope. As seen in Fig. 4, a decrease of ϳ60% in intracellular Salmonella was observed 2 h postinfection in the R37 cell line, which stably expresses functional Nramp1. As expected, no killing of intracellular pathogen was observed in R21 cells, which expresses Nramp1 antisense. This fits previous observations that Nramp1 is crucial for macrophage resistance to S. enterica serovar typhimurium (30).
Peritoneal Macrophages Derived from IRF-8 Ϫ/Ϫ Mice Exhibit Impaired Killing of S. enterica Serovar typhimurium-To find a link between IRF-8 expression in macrophages and resistance to S. enterica serovar typhimurium, we used peritoneal macrophages, which are considered as professional macrophages. These cells were collected from both WT and IRF-8 Ϫ/Ϫ mice  and following activation with IFN-␥ and LPS were infected with S. enterica serovar typhimurium and killing assay was performed as described under "Experimental Procedures." Intracellular S. enterica serovar typhimurium viable counts were decreased by more than 10-fold 2 h postinfection and remained at this low level until the experiment was terminated (Fig. 5A,  solid line). On the other hand, only a ϳ0.5 log decrease in the intracellular S. enterica serovar typhimurium viable counts was observed for peritoneal macrophages derived from IRF-8 Ϫ/Ϫ mice (Fig. 5A, dashed line).
To evaluate the efficiency of phagocytosis and the functioning of the developing phagosome, the cells were infected with S. typhimurium expressing GFP and following 45 min of invasion the cells were stained with LysoTracker Red (Cambrex), which stains acidic compartments in living cells, and 4Ј,6-diamidino-2-phenylindole, which stains the nuclei. The cells were fixed and observed under fluorescent microscope. It is clear that both types of peritoneal macrophages equally phagocytized GFP expressing S. enterica serovar typhimurium, however, the Lyso-Tracker red staining of phagosomes of IRF-8 Ϫ/Ϫ macrophages was significantly weaker, indicating that the phagosome is less acidified (Fig. 5B). This may be due to lack of Nramp1 expres-sion leading to a significant reduction in the acidity of the phagosome as reported previously (31)(32)(33). In these experiments the IRF-8 Ϫ/Ϫ mouse strain (129/Sv IRF-8 Ϫ/Ϫ ) was wild type for Nramp1 as will be elaborated hereafter.

Loss of Either IRF-8 or Nramp1 Results in an Aberrant Killing Response of Macrophage against S. enterica Serovar typhimurium and M. bovis (BCG)-
Our studies thus far assign a pivotal role for IRF-8 in Nramp1 regulation and in innate resistance to intraphagosomal pathogens such as S. enterica serovar typhimurium. We next wanted to directly link the defective bactericidal response of IRF-8 Ϫ/Ϫ macrophages with aberrant expression of Nramp1. The original IRF-8 knock-out mouse strain was generated in C57BL/J6, which also have a nonfunctional Nramp1 allele due to a single amino acid mutation (Asn-169) (5). Therefore, to genetically dissociate between the effect of IRF-8 and Nramp1 on the ability of macrophage to eliminate intraphagosomal pathogens, we had to generate a new mouse strain by cross-breeding for 8 generations the mouse strain 129/Sv harboring endogenous functional Nramp1 and the C57BL/J6 mouse strain ablated for IRF-8. A comparative S. enterica serovar typhimurium killing study was performed with 4 different mouse strains. The two wild type mouse strains, C57BL/J6, which is practically knock-out for Nramp1; IRF-8 ϩ/ϩ Nramp1 Ϫ/Ϫ ) and 129/Sv that harbors intact genes. In addition, 129/Sv IRF-8 Ϫ/Ϫ (Nramp1 ϩ/ϩ ) and C57BL/J6 IRF-8 Ϫ/Ϫ , which is also Nramp1 defective and thus practically a double knock-out mouse strain. Peritoneal macrophages were collected from these four mouse strains and following 18 h of activation with IFN-␥ and LPS, the cells were infected with S. enterica serovar typhimurium. As seen in Fig. 6A, the killing of S. enterica serovar typhimurium by the wild type 129/Sv mouse strain was the most effective. However, the killing ability of macrophages that are defective for IRF-8, Nramp1, or both was identical and markedly impaired compared with the wild type macrophages.
Similarly, bone marrow macrophages were collected from these 4 mouse strains and after 6 days of in vitro cultivation the cells were activated with IFN-␥ and LPS for 18 h and then infected with M. bovis (BCG) as described under "Experimental Procedures." Similar to the data observed in Fig. 6A, WT macrophages killed this intraphagosomal pathogen more effectively than macrophages that are defective for IRF-8, Nramp1, or both, which exhibited an almost identical killing pattern of M. bovis (BCG) (Fig. 6B). Linking mouse genetics with macrophages phenotypic susceptibility to intraphagosomal infection our results strongly suggests that IRF-8 is a major regulator of Nramp1, enabling an efficient elimination of these pathogens at early stages of infection.
To directly show that IRF-8 is an immediate regulator of Nramp1 in vivo, bone marrow macrophages were collected from the 129/Sv IRF-8 Ϫ/Ϫ mouse strain and transduced with either empty pMSCV retroviral vector or with the same vector harboring IRF-8. After 7 days, the cells were activated with IFN-␥ and LPS and subjected to killing assay as described under "Experimental Procedures." It is clear from Fig. 6C that rescuing IRF-8 expression led to a significant improvement in S. enterica serovar typhimurium killing profile. Taken together, our results demonstrate that ablation of either IRF-8 or  Fig. 4. B, to determine phagocytic activity and phagosome acidity, cells were seeded on coverslips and challenged with the pathogen (expressing GFP) as described above. At time 0, phagosomes were stained with Lyso-Tracker Red, nuclei with 4Ј,6-diamidino-2-phenylindole, and phagocytized Salmonella were observed due to the expression of GFP as described under "Experimental Procedures." Cells were fixed and slides were mounted and observed under a fluorescent microscope.
Nramp1 or both lead to the same killing activity in macrophages. Additionally, rescuing IRF-8 expression improved the killing ability of IRF-8 Ϫ/Ϫ cells that have a WT Nramp1 allele.
These data suggest a direct genetic link between IRF-8 expression, Nramp1 regulation, and phagosome-mediated elimination of intraphagosomal pathogens such as S. enterica serovar typhimurium or M. bovis (BCG) at the early stage after infection.

DISCUSSION
Innate resistance to intraphagosomal pathogens is mediated by various genes enabling the elimination of the invader. Nramp1 was identified as one of the major constituents that confer mice resistance to such pathogens. In humans, aberrant expression of NRAMP1 is associated with susceptibility to infectious and autoimmune disease (34,35).
Nramp1 is expressed exclusively in myeloid cells such as macrophages and dendritic cells (36 -38). Our previous study demonstrated that synergistic activation of Nramp1 is mediated through association of Miz-1, a ubiquitously expressed factor, with the hematopoietic specific transcription factors IRF-8 and PU.1 and laid the molecular basis for the cell type restricted expression of Nramp1. PU.1 is constitutively expressed at relatively high levels in monocyte/macrophage cells and cannot account for the IFN-␥ inducibility of Nramp1 (11). IRF-8, on the other hand, is induced by IFN-␥, and serves as the catalyst for the formation of the three party heterocomplex. The interaction between IRF-8 and PU.1 is a paradigm for macrophage-specific expression of many genes. Following protein-protein interaction, the heterocomplex is capable of binding to a DNA composite element of which half is the IRF binding site and half is the PU.1 binding site (also termed as ETS) (10,39). Here we show that the binding sites for IRF-8 and PU.1 are separated by ϳ32 bp, however, protein-protein interaction is essential because mutation in the IRF association domain of IRF-8 eliminates the synergistic interaction with Miz-1 and the further enhancement with PU.1 (6). The region just upstream of the ISRE and PU.1 binding site is characterized by 27 ϫ GT repeat in mice that is partially conserved in human. In human, the segmented GT repeat is subjected to polymorphism that affects Nramp1 promoter activity (40,41). This GT repeat can assemble in a condensed Z-DNA structure in vivo (40) that may lead to juxtaposition of the two binding sites and facilitate heterocomplex formation.
Previous studies suggested that the transcriptional enhancement elicited by Miz-1 is due to interaction with the co-activator p300. This is repressed by c-Myc, at least in part, through competition for binding with p300 (42). Here we show that this repression is also due to c-Myc interfering with Miz-1 and IRF-8 interaction. Miz-1 is a multidomain transcription factor and its interaction with c-Myc is via helix-loop-helix domain that lies between amino acids 683 and 715 (11,27). The interaction with IRF-8 is mediated through a PEST domain that spans amino acids 190 -273 (6). This fits with previous studies demonstrating that IRF-8 association with other transcription factors is mediated through IRF association domain-PEST domain interaction. Therefore, it suggests that Miz-1 interacts with c-Myc and IRF-8 through different domains that are probably sequestered by the counterinteracting partner: c-Myc or IRF-8. We propose that in resting macrophages c-Myc levels are sufficient to repress Miz-1 activity by blocking possible interaction with p300 and IRF-8. When the cells are activated by IFN-␥ or pathogens, IRF-8 is induced and concomitant c-Myc expression is down-regulated as previously reported (43). Moreover, it has been recently shown that IRF-8 induction leads to c-Myc repression (44). Therefore, in activated macrophages, IRF-8 competes with c-Myc for the association with Miz-1. Upon IRF-8 interaction, PU.1 is also assembled leading to maximal activation of Nramp1 expression. This cell typespecific activation of Nramp1 by IRF-8 and PU.1 also occurs during infection with S. enterica serovar typhimurium and L. donovani as shown here. Taken together, this suggests that during macrophage activation the repression of Nramp1 due to Miz-1/c-Myc interaction is alleviated and c-Myc is being replaced by IRF-8/PU.1 binding that lead to the assembly of a cell type-specific enhanceosome. This signaling cascade is turned off due to the negative feedback loop that deactivate IFN-␥ signaling and concomitantly lead to restoration of c-Myc expression and down-regulation of IRF-8 expression. Consequently, the enhanceosome architecture returns to the steady state status.
The importance of Nramp1 to the killing of intraphagosomal pathogens is well established both in vitro and in vivo (45,46). In this article, we have established the role of IRF-8 as an important regulator of Nramp1. To do that, we have generated a new IRF-8 Ϫ/Ϫ mouse strain with a genetic background containing a functional Nramp1 allele. Macrophages from all 4 possible combinations of WT and mutated genes for IRF-8 and Nramp1 were studied. Infection of these cells with S. enterica serovar typhimurium or M. bovis (BCG) clearly demonstrated that ablation of either IRF-8 or Nramp1 led to a similar deficiency in killing activity following pathogen invasion as also seen with macrophages retrieved from the double ablated/mutated mouse strain. The fact that the killing efficiency of the IRF-8 null strain resembled that of the Nramp1 defective mouse strain and the double knock-out strain suggests that these two genes are phenotypically as well as genetically associated. Together with our Nramp1 promoter analysis we conclude that in vivo IRF-8 also controls the expression of Nramp1. To further establish this, we have complemented bone marrow-derived macrophages from IRF-8 Ϫ/Ϫ /Nramp1 ϩ/ϩ mice with retroviral vector allowing IRF-8 expression. This led to partial recovery of the killing ability that was statistically significant. This killing activity was better than the killing activity observed for the double knock-out macrophages complemented with just IRF-8 (data not shown). These results point to the pivotal role of IRF-8 as a transcriptional regulator of the expression of Nramp1 in vivo. This is in line with previous publications describing the sensitivity of IRF-8 null mice to intraphagosomal pathogens such as Leishmania and Toxoplasma (47,48). IRF-8 is a key factor not only for monocyte/macrophage differentiation but also to the function of mature macrophages. As such, it is a major regulator of IL-12p40 and the sensitivity to the various pathogens was solely attributed to lack of Th1 response in the IRF-8 knock-out mouse strain due to lack of IL-12 expression. Later, it was demonstrated that IRF-8 is also a key regulator of phagosome essential genes (phagocytic oxidase (12), inducible nitric-oxide synthetase (13)), inflammatory cytokine (16), and Toll receptors 4 and 9 (49,50), which all are essential for macrophage-mediated killing of pathogens by affecting both innate and adaptive immune responses. Naturally, it is also involved indirectly being part of the signaling cascade of Toll receptors and IFNs (22,51). To this list of genes regulated by IRF-8, which are essential for intraphagosomal pathogen elimination, we add Nramp1.
A recent publication (21) looking at BXH-2 mice harboring a point mutation in the IRF association domain of IRF-8 (R294C) demonstrated that these mice were also susceptible to S. enterica serovar typhimurium to a level comparable with that seen for mice lacking functional Nramp1 or TLR4. In addition, these mice were permissive to M. bovis (BCG), despite a resistant Nramp1G169 allele, and were unable to control splenic bacterial replication, which continued for up to 8 weeks postinfection. These impaired IRF-8 R294C mice also demonstrated increased replication of the Plasmodium chabaudi AS malarial parasite during the first burst of blood parasitemia, and recurring waves of high blood parasitemia late during infection (21). The conclusions were that IRF-8 is required for orchestrating early innate responses and also long-term immune protection against unrelated intracellular pathogens. In general, these results are in line with our results. However, the results described here are related only to the early stages of macrophage infection in vitro. Under these experimental conditions, immune components such as adaptive immunity are not studied.
To conclude, in this article we show that the diverse effects of IRF-8 on innate immunity are also funneled through macrophage-restricted regulation of Nramp1. The Nramp1 enhanceosome is subjected to a delicate balance between a repressor (c-Myc) during normal physiological condition and an activator (IRF-8) induced in response to infection with intraphagosomal pathogens. This interplay between repression and activation explain in molecular and genetic terms the ability of macrophages to eliminate such pathogens at early stages of infection.