The Far Upstream Chicken Lysozyme Enhancer at -6.1 Kilobase, by Interacting with NF-M, Mediates Lipopolysaccharide-induced Expression of the Chicken Lysozyme Gene in Chicken ~-Myelomonocytic Cells*

Macrophages respond to lipopolysaccharide (LPS) with the activation of various genes, including the lysozyme gene. Here, we show that the level of lysozyme mRNA increases following treatment of chicken myelomonocytic HDll cells with LPS. By transient and stable transfection of the chloramphenicol acetyltrans-ferase (CAT) gene controlled by regulatory elements of the lysozyme gene, we identified a subfragment of the -6.1 kilobase (kb) lysozyme enhancer that mediates the LPS-induced lysozyme expression. This subfragment contains two elements (D and E), each of which matches the highly degenerate consensus sequence of binding sites for CIEBP-like transcription factors. Furthermore, we found protein complexes to interact with elements D and E whose binding activity to elements D and E is LPS-inducible in myelomonocytic HDll cells. Immunomobility shift assays show that NF”M, a myeloid-specific CIEBPP-related transcription factor is an essential component of these protein complexes. Mutations of the C/EBP binding sites within D and E cause a reduction of basal activity and abolish LPS responsiveness of the -6.1 kb lysozyme enhancer. These results show that the

Lysozyme is one of the antibacterial proteins produced in chicken oviduct and macrophages. The chicken lysozyme gene provides an attractive model for studying the cell-and cell stage-specific expression of eukaryotic genes. In tubular gland cells of the chicken oviduct, the transcriptional activity of the lysozyme gene is strictly dependent on the development of the oviduct. Thus, expression of the lysozyme gene in the oviduct is regulated directly by steroid hormones (1). The chicken lysozyme gene is also a later marker gene for the myeloid lineage of hematopoietic differentiation. In mature macrophages, it is constitutively expressed and is independent of steroid hormones (2). During the differentiation of macrophages, the lysozyme gene is continuously activated with low expression in myeloid cells and high expression in mature macrophages. The transcriptionally active lysozyme gene is located in a chromatin domain displaying generally higher nuclease sensitivity, which * This work was supported in part by a grant from the Deutsche Forschungsgemeinschaft (to L. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "uduertisenent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom all correspondence and requests for reprints should be addressed. Tel.: 5141-384646; Fax: 5141-381849. is approximately 19-24 kb' in size (3, 4). The 5' and 3' borders of this nuclease-sensitive domain coincide with the nuclear matrix attachment regions (5-9). All known transcriptional regulatory elements involved in the tissue-specific and developmentally regulated expression of the gene are contained in this chromatin domain. Glucocorticoid and progesterone responsive elements are located at -0.2 kb and -1.9 kb (10,11). In addition, cell type-specific transcriptional enhancers have been identified at -0.2 kb, -2.7 kb, and -6.1 kb (12)(13)(14). Whereas the far upstream enhancer at -6.1 kb is macrophage-and oviductspecific, the -0.2 kb and the -2.7 kb enhancer are only active in cells of the myeloid lineage in the hematopoietic system of the chicken.
LPS, in the acute phase of bacterial infection, activates important cellular mechanisms modulating immunological and inflammatory responses. LPS has been shown to elevate gene expression. The activation is regulated at the transcriptional level by transcription factors such as NF-KB arm-1 (15,161. In order to study the activation of macrophages by LPS, we determined the effect of LPS on lysozyme gene expression in chicken myelomonocytic HDll cells. We found that LPS causes dramatically morphological changes of HDll cells correlating with the increase of lysozyme mRNA level. LPS activates transcriptional activity of elements D and E of the lysozyme enhancer at -6.1 kb. Furthermore, we demonstrate that NF-M, a myeloid-specific C/EBP-like transcription factor, is present in the LPS-inducible protein complexes that bind to the C/EBP binding sites of elements D and E, and this binding is the cause for transcriptional activation of the -6.1 kb lysozyme enhancer by LPS. Our results show that the -6.1 kb lysozyme enhancer, in addition to its role in cell type-specific lysozyme gene expression during differentiation of macrophages, by interacting with C/EBP-like transcription factors can modulate the lysozyme expression in response to external stimuli that activate macrophages.

MATERIALS AND METHODS
ard procedures (17). To construct pcEPCAT5, plasmid pLYSCAT2100 DNA Constructs-Plasmid constructions were performed by stand- (13) was digested with Hind111 and BgZII, the 1.2-kb HindIII-BgZII fragment containing the chicken lysozyme promoter (-579/+14), and the -6.1 kb lysozyme enhancer (-6331/-5769) was inserted into pBLCAT5 cut previously with Hind111 and BumHI. pcEPml, in which the CeZII-AccI fragment is deleted, was constructed as follows. The pcEPCAT5 was digested with CeZII, and the digested ends were made acetyltransferase; C/EBP, CCAAT enhancer binding protein; CRH, cor-The abbreviations used are: kb, kilobase(s); CAT, chloramphenicol ticotropin-releasing hormone; cMGF, chicken myelomonocytic growth factor; HIV-I, human immunodeficiency virus, type I; LPS, lipopolysaccharide; NF-M, myeloid-specific transcription factor; NF-KB, nuclear factor KB; Me,SO, dimethyl sulfoxide. flush with Klenow DNA polymerase and subsequently cut with HindIII. The CelII-Hind111 fragment was ligated into pcEPCAT5, which was previously cut with A d , and then blunted by Klenow DNApolymerase, and finally cut with HindIII. The pcEPm2, in which the NcoI-CelII fragment is deleted, was constructed from pcEPCAT5 in a similar way with the exception that NcoI and CelII were used.
pcPCAT was constructed by insertion of the Sau3A fragment from pcEPCAT5 containing the lysozyme promoter (-579/+14) in the BglII site of the pBLCAT5. p-6.3/+1.8CAT containing the B2X2 fragment (-6.3/+1.8) of the lysozyme gene upstream of the CAT gene was constructed as follows. The HindIII-XhoI sequence carrying the multiple cloning sites of the pBLCAT5 was substituted by a synthetic oligonucleotide that contains an internal BamHI site upstream of an XbaI site. After subcloning, the B2X2 fragment was inserted into the BamHI, and XbaI sites generated the p-6.3/+1.8CAT.
The plasmid pCRH containing the CAT gene driven by the human corticotropin-releasing hormone (CRH) promoter, was described previously (18). p-6.lCRH was constructed by insertion of the -6.1 kb lysozyme enhancer in front of the CRH promoter.
p-2.7CRH was constructed in a similar way by using the -2.7 kb myeloid-specific lysozyme enhancer.
To generate oligonucleotide-CRH promoter constructs containing different elements (T2, C, and D-E) of the -6.1 kb lysozyme enhancer (14), annealed oligonucleotides containing 5"overhang sequences for HindIII and BamHI restriction sites were purified by polyacrylamide gel electrophoresis and cloned into the HindIII and BamHI sites of pCRH.
pSV-Gal from Promega (Heidelberg, Germany) was used as an internal control for transfection efficiency.
Oligonucleotides-Synthetic oligonucleotides (5'-end position is given before sequence, and altered bases are underlined) were purchased from MWG-Biotech (Ebersberg, Germany) and from Pharmacia Cell Culture and DNA Dansfections-HD11 cells (19) were maintained in Iscove's modified Dulbecco's medium, supplemented with 8% fetal calf serum, and 2% chicken serum, 100 unitslml penicillin, and 100 pg/ml streptomycin at 37 "C and 5% CO,. HD57 cells were maintained as described (20). Transfections of HDll cells were performed as described elsewhere (18). Transfected cells were fed with 10 ml of fresh medium and incubated, when indicated, with 5 pg/ml LPS from Salmonella typhimurium (Sigma, Germany) at 37 "C and 5% CO,. Twentyfour hours after transfection, transfected cells were harvested for preparation of protein extracts and determination of CAT and P-galactosidase activity. Protein concentrations of cell extracts were determined by the method of Bradford (Bio-Rad Laboratories, Munich, Germany). For stable transfection, 20 pg pcEPCAT5 and 2 pg ptkNeo were used, transfected cells were fed with fresh medium and incubated for 24 h at 37 "C before 500 pg/ml of G418 was added to select for cells that had stably integrated DNA. G418 resistance clones were isolated 2-3 weeks later and expanded for induction experiments with LPS.
CAT and P-Galactosidase Assay-CAT activities were determined with 200 pg of total protein of cell extracts for transient expression assays and 50 pg of total protein for stable expression assays (18). After autoradiographic exposure, unmodified chloramphenicol and its acetylated derivates were excised from thin-layer chromatography plates and counted in a liquid scintillation counter. Assays for p-galactosidase were performed by the standard method (17). Results are representative of three independent transfections and were normalized to P-galactosidase activities. Northern analysis was performed by the standard method (17). Briefly, poly(A)+ RNAs were denaturated with glyoxal and dimethyl sulfoxide, fractionated on a 1.5% agarose gel, transferred to nylon membrane (Appligene), and baked at 80 "C for 2 h. The plasmid pls-1 (21) containing the lysozyme cDNA was labeled by nick translation (Life Technologies, Inc. kit). Hybridization was performed according to the method described by Church and Gilbert (22). After washing, the filter was dried and exposed to x-ray film.
Lysozyme and glyceraldehyde-3-phosphate dehydrogenase mRNA levels were quantitated by an RNase protection assay according to the method described by Melton et al. (23). Labeled antisense RNAs were obtained as follows. The TaqI-KpnI fragment (252 base pairs long) of lysozyme cDNA (24) and the HindIII-PstI fragment of glyceraldehyde-3-phosphate dehydrogenase cDNA (25) were cloned into pBluescript I1 SK (Stratagene). The recombinant plasmids were linearized either with XbaI for lysozyme cDNA or with XhoI for glyceraldehyde-3-phosphate dehydrogenase cDNA and used as templates to synthesize singlestranded antisense RNAs using T7 or T3 RNA polymerase in a transcription assay containing [CY-~~PIUTP (Stratagene kit). RNase protection was performed by hybridizing in solution 4 pg of poly(A)+ RNA and antisense 32P-labeled RNA probe for lysozyme and glyceraldehyde-3phosphate dehydrogenase as control and digesting with RNase T1 and RNase A (17). Protected RNAs were fractionated on 5% polyacrylamide, 8 M urea gels. After drying, gels were exposed to x-ray films.
Nuclear Extract Preparation and Electrophoretic Mobility Shift Assay (EMSAI-Nuclear extracts from exponentially growing cells were prepared by a rapid method according to Schreiber et al. (26).

Double-stranded oligonucleotides were labeled with [CP~~PI~CTP and
Klenow DNA polymerase by fill-in reaction. Band shift assays were performed in a total volume of 20 pl containing 0.5 pg of poly(d1-dC) (Pharmacia, Germany), 5 pg of protein of nuclear extracts, 3 m M HEPES-KOH, pH 7.9, 2.4 m M EDTA, 1 m M dithiothreitol, 2 m M spermidine, 10 mg/ml bovine serum albumin, 120 m M NaCl, 4% Ficoll 400, and 20,000 cpm of labeled double-stranded oligonucleotide. Nuclear extracts were preincubated with poly(d1-dC) or, when indicated, with specific or unspecific competitors at 4 "C for 5-10 min before incubation with labeled probe at 4 "C for an additional 30 min. After that, samples were directly loaded on a 4% non-denaturing polyacrylamide gel in 0.25 x TBE. The gel was run at room temperature, at 150 V for 1.5 h, fixed in 10% acetic acid, dried, and autoradiographed at -80 "C. For immunomobility shift assays, nuclear extract was incubated with labeled oligonucleotide for 15 min at 4 "C, diluted antiserum or preimmune serum was then added for an additional 15 min before loading on a gel.
Phosphatase Deatment of Nuclear Extracts-Nuclear extracts of HDll cells were treated with acid phosphatase according to Williams and Maizels (27).

LPS Induces Morphological Changes of Chicken Myelomonocytic Cells and Expression of the Chicken Lysozyme Gene-LPS is an integral component of the outer membrane of Gram-negative bacteria. It has multiple effects on macrophages such as induction of cytotoxic activity. Here, we examined effects of LPS on morphology of HDll cells of a chicken myelomonocytic line transformed by the u-myc-encoding MC29 virus (19) and on expression of the lysozyme gene. Cells treated with LPS
showed several characteristics of mature macrophages, they became tightly adherent, were increased in size, and displayed a cytoplasm containing numerous granules and large vacuoles, and developed numerous lamellipodia. able only in RNase protection assays (Fig. LA). Upon the treatment of LPS, the levels of lysozyme mRNA in H D l l cells were rapidly induced within the first 30 min of incubation and continued to accumulate and remained high until a t least 10 h after LPS addition to HDll cells (Fig. LA). In contrast, the level of glyceraldehyde-3-phosphate dehydrogenase mRNA was not affected by LPS and remained constant during the induction by LPS (Fig. 1 B ) . Similar results were obtained by a Northern analysis. However, because of lower sensitivity of Northern analysis, no detectable lysozyme transcripts were found in control cells and in cells treated with LPS for 30 min (Fig. 1C).
LPS Induces Dansient Expression Controlled by the 5'-Flanking Region of the Lysozyme Gene-To examine whether the DNA sequence upstream of the coding region mediates induction of expression of the lysozyme gene by LPS, we transfected the plasmids pcPCAT, controlled by the lysozyme promoter alone, and p-6.3/+1.8CAT, driven by the 5"flanking region (-6.3/+1.8) and part of the coding sequence of the lysozyme gene into H D l l cells. For controls, we used the CAT reporter gene construct under control of the thymidine kinase promoter from herpes simplex virus (pBLCAT4). Transfected cells were subsequently treated with 5 pg/ml LPS. As shown in Fig. 2, expression of pcPCAT is increased 6-fold in LPS-stimulated H D l l cells. Treatment with LPS, however, strongly induces CAT activity from pcEPCAT5 and p-6.3/+1.8CAT (15.2-fold and 16.8-fold, respectively). The low CAT activity of the plasmid p-6.3/+1.8CAT in HDll cells may be explained by a low transfection efficiency, because this plasmid is 12.4 kb in size. In contrast, expression of the pBLCAT4 is 1.9-fold inducible by LPS. Our data show that the 5'-flanking region of the lysozyme gene contains regulatory cis-elements that mediate LPS-induced gene expression.
The chicken lysozyme promoter contains a &-like sequence (-164/-155). This element can mediate NF-KB-regulated expression of the lysozyme gene? To determine whether the KBlike sequence is essential for the LPS-induced expression of the lysozyme gene, we performed transient expression assays with the pcEPCAT5 containing the lysozyme promoter (-579/+14) and the -6.1 kb lysozyme enhancer and two deletion mutants of the lysozyme promoter (Fig. 3A). As shown in Fig. 3B, expression of the CAT gene from these plasmids as from pcEPCAT5 in HDll cells was strongly induced by LPS to the same extent. These results indicated that other cis-elements than the KBlike sequence are dominantly required for the induction of lysozyme gene expression by LPS. A 44-Base Pair Fragment of the -6.1 kb Lysozyme Enhancer Mediates Activated Expression by LPS-In order to define regulatory elements of the lysozyme gene required for induction by LPS, we transiently transfected the p-2.7CRH and p-6.lCRH into HDll cells in the presence of LPS. The parental plasmid pCRH was used for controls. The CAT gene controlled by the human CRH promoter has been shown previously to be expressed in chicken H D l l cells (18). When compared with basal activity of pCRH, the -6.1 kb lysozyme enhancer stimulates .T expression only 2.5-fold in HDll cells, whereas the -2.7 kb lysozyme enhancer confers stronger transcriptional activity, activating expression of the adjacent CRH-CAT gene 96-fold in the HDll cell line (Fig. 4A). Fig. 4A also shows that expression of the CAT gene driven by the CRH promoter was not inducible by LPS. The -2.7 kb lysozyme enhancer mediates 1.8-fold LPS induction. This weak activation may be caused by the presence of an AP-1-like sequence in the -2.7 kb lysozyme enhancer (28). In contrast, in the presence of the lysozyme enhancer at -6.1 kb, expression of the CAT gene under control of the CRH promoter was clearly induced (5.3-fold) by LPS. These results show that the -6.1 kb lysozyme enhancer contains regulatory elements that are important for the LPS-induced lysozyme gene expression.
The -6.1 kb lysozyme enhancer contains multiple elements that contribute to its stimulatory function (14). Further, to define minimal sequences sufficient for LPS-activated expression, we cloned elements T2 (-6010/-5991), C (-5989/-5970), D (-5967/-5950), and E (-5947/-5930) of the -6.1 kb lysozyme enhancer upstream of the human CRH promoter. Because there are some palindromic sequences spanning the D and E elements, we cloned these elements together into the plasmid pCRH. The ability of these chimeric constructs to direct LPSinducible CAT expression in HDll cells was then assayed. As shown in Fig. 4B the element T2 containing a palindromic NF-1 motif imparted no LPS-inducible activity, whereas the element C containing an AP-1-like sequence (14) activated CAT activity about 2.2-fold after stimulation with LPS. In contrast, the downstream elements D and E elevated CAT expression 12-fold in unstimulated HDll cells and significantly mediated LPSinduced expression of the CAT gene (6.1-fold).
LPS Activates Expression of the Lysozyme-CAT Construct Stably Integrated in HDll Cells-To analyze the effect of LPS on expression of the lysozyme-CAT construct, when it is stably integrated into genome, we co-transfected the plasmid pcEP-CAT5 together with a neomycin resistance gene plasmid into HDll cells.

FIG. 2. The 5'-flanking region of the lysozyme gene mediates LPS-induced expression.
HDll cells were transfected with pcPCAT, pcEPCAT5, or p-6.3/+1.8CAT or with pBLCAT4 for control.  elements D and E of the -6.1 kb lysozyme enhancer mediate LPS-activated expression from the human CRH promoter. To gain insight into the mechanism by which the -6.1 kb lysozyme enhancer contributes to LPS-inducible expression of the lysozyme gene, we assayed for LPS-inducible DNA binding proteins that bind to these elements. We performed EMSAs with

FIG. 4. LF'S stimulates transcriptional activity of elements D and E of the -6.1 kb chicken lysozyme enhancer. A, HDll cells
were transfected with pCRH, p-6.lCRH containing the -6.1 kb lysozyme enhancer, and p-2.7CRH containing the -2.7 kb lysozyme enhancer. B, HDll cells were transfected with CRH-CAT constructs containing the subfragments of the -6.1 kb lysozyme enhancer. The basal CAT activity from pCRH treated with Me,SO (DMSO) was assigned a value of 1.0. The relative CAT activities were mean values from three independent transfection experiments. Standard deviations were less than 20%.
double-stranded oligonucleotides containing either element D or E of the -6.1 kb lysozyme enhancer. As shown in Fig. 6 A , nuclear extracts of H D l l cells treated with LPS for 10 h connuclear extracts from H D l l cells. Radiolabeled probes were tained DNA binding proteins, which bind to radiolabeled oligo-  SO (lunes 1,  3, 5, 7, and 9) for control or with LPS  (lanes 2, 4, 6, 8, and 10). nucleotides D and E. To demonstrate the sequence specifity of these DNA-protein complexes, we performed competition assays in the presence of unlabeled oligonucleotides. Oligonucleotides D and E competed with each other for binding of nuclear factors, implying that factors binding to D and E have similar binding specificities. In contrast, unlabeled oligonucleotides containing elements T2 and C of the -6.1 kb lysozyme enhancer cannot compete for the D and E elements even a t a 200-fold molar excess. Next, to determine whether nuclear factors that bind to elements D and E are LPS-inducible, we performed binding assays with nuclear extracts from H D l l cells treated with LPS for up to 10 h. As shown in Fig. 6B, a nuclear extract of HDll cells treated with LPS for 2 h contained already enhanced binding activity to oligonucleotides D and E. This binding activity increased time dependently in H D l l cells treated with LPS for up to 10 h.

LPS-inducible Factors Binding to Elements D and E are Myeloid-specific and Compose NF-M, a Myeloid-specific Dunscription Factor-
To determine whether nuclear factors binding to the D and E elements are expressed and inducible in other cells, we prepared nuclear extracts from chicken erythroid HD57 cells (201, from chicken embryo cells, and from DU249 cells of a chicken liver cell line (29) for binding assays with labeled D oligonucleotide. As shown in Fig. 7A, LPS-inducible nuclear factors binding to element D are present in nuclear extracts of myelomonocytic H D l l cells but not in nuclear extracts of HD57 cells, embryonic cells, and DU249 cells, even though these cells were treated with LPS for 10 h.
A sequence analysis of the -6.1 kb lysozyme enhancer re-  (Fig. 7B). Fig.  7C shows that, like the unlabeled E oligonucleotide, only oligonucleotides containing a C/EBP binding site competed well For the E oligonucleotide. In contrast, the KB motif of HIV-I and its mutant mKB motif are not able to inhibit binding activity to the D and E elements. The results suggest that nuclear factors that bind to elements D and E of the -6.1 kb lysozyme enhancer are related to transcription factors of the C/EBP family.
To identify the C/EBP binding activity in H D l l cells, immunoshifts were performed with an antiserum raised against NF-M. NF-M, a myeloid-specific C/EBP-like transcription factor, should play a role in signal transduction and differentiation of avian myelomonocytic cells (34). As shown in Fig. 8, A and B, only the antiserum to NF-M reacted specifically with the complexes formed between H D l l nuclear extract and D or E. Low concentrations of the NF-"specific antiserum caused new complexes with lower mobility, and higher concentrations of the cleotide was ineffective (Fig. 9A). Next, to assay the LPS inducibility driven by these oligonucleotides, mD-E and mD-mE E were then cloned into the CRH promoter-CAT construct and 1 I transfected into HDll cells. Fig. 9B shows that, when com- indicating that C/EBP binding sites are essential for basal activity of the -6.1 kb lysozyme enhancer in myelomonocytic cells. Furthermore, mD-mE failed to impart LPS-inducible activity, whereas mD-E shows reduced LPS responsiveness by mutation of element D. Taken together, these data correlate with binding activities of NF-M to oligonucleotides D-E, mD-E, and mD-mE, implying that at least two intact C/EBP binding sites of the -6.1 kb lysozyme enhancer are required for fully LPS-activated expression of the lysozyme gene in myelomonocytic cells. NF-"specific antiserum eliminated over 90% of HD11-specific complexes. No effects were observed with preimmune serum. Because all complexes were equally affected by NF-"specific antiserum, we suggest that NF-M is an essential component in all of these complexes that bind to elements D and E of the -6.1 kb lysozyme enhancer.

Interaction of NF-M with Elements D and E is Required for the Cell-specific and LPS-activated
Lysozyme Gene Expression-To determine whether the observed LPS activation correlates with an interaction of NF-M with elements D and E, we had oligonucleotides with point mutations within D (mD-E) or within D and E (mD-mE) synthesized. mD-E oligonucleotide containing one intact C/EBP binding site was able to compete effectively for E oligonucleotide, but mD-mE oligonu-domain adjacent to a basic DNA binding region and can form homodimers or heterodimers with other bZip proteins (35,36). Transcriptional activation by these factors can be enhanced by phosphorylation near and within the bZip region (37,38). To determine whether binding activity to D and E depends on phosphorylation, we performed binding assays in which 32Plabeled synthetic oligonucleotide E was incubated with nuclear extracts treated briefly with acid phosphatase. Fig. 10 shows that binding activity of protein complexes in untreated HDll cells was sensitive to treatment with acid phosphatase. However, binding activity in LPS-stimulated cells was only marginally affected by phosphatase treatment. Therefore, this result indicates that LPS-enhanced binding activity to D and E of the -6.1 kb lysozyme enhancer does not seem to be a result of phosphorylation of protein complexes in nuclei of HDll cells. sozyme mRNA correlated temporally with the NF-M activation. However, because the stimulation of lysozyme mRNA seems to be more prominent than the NF-M activation by LPS, we cannot rule out posttranscriptional activation mechanisms involved in the LPS activation of the lysozyme gene. We were interested in underlying molecular mechanism of the lysozyme induction by LPS. Our data show that the -6.1 kb lysozyme enhancer can mediate the LPS induction of lysozyme gene expression. Surprisingly, the -6.1 kb lysozyme enhancer with a length of 562 base pairs stimulated the CAT expression driven by the human CRH promoter less efficiently than its smaller subfragments (T2 and D-E). This effect may be caused by different sizes of the fragments cloned upstream of the human CRH promoter and resulted in different transfection effkiencies. The -6.1 kb lysozyme enhancer, as the &-like sequence of the lysozyme promoter, is also able to mediate LPS induction. Although the -6.1 kb enhancer yields a 15.2-fold LPS stimulation in expression from the lysozyme promoter and only a 5.3fold increase in expression from the CRH promoter, it is unlikely that the &-like sequence of the lysozyme promoter and the -6.1 kb lysozyme enhancer act cooperatively in the activation by LPS, because a deletion of the CeZII-AccI fragment carrying the &-like sequence did not lead to a partial loss of response of the pcEPml to LPS. The difference in the extent of LPS activation (15.2-fold versus 5.3-fold) may be explained by assuming that the -6.1 kb lysozyme enhancer with its binding proteins together with the lysozyme promoter works more effectively with the transcriptional machinery than in the CRH promoter context. We also suppose that the CRH promoter may contain sequences that interfere with the -6.1 kb lysozyme enhancer, resulting in less LPS responsiveness.
We were interested to learn about the molecular mechanism by which regulatory elements of the -6.1 kb lysozyme enhancer and their binding proteins contribute to LPS-activated expression of the lysozyme gene. Our studies have demonstrated that LPS-induced expression of the lysozyme gene in myelomonocytic cells is, to a large extent, the result of interaction of LPSinducible transcription factors with C/EBP binding sites that are located within elements D and E of the -6.1 kb lysozyme enhancer. Competition experiments with oligonucleotides containing binding sites of C/EBP transcription factors indicate that transcription factors binding to elements D and E belong to the C/EBP family. In fact, immunomobility shift assays with a specific antibody to NF-M have shown that NF-M is present in all of the protein complexes binding to the elements D and E. NF-M would be the chicken homologue of the human NF-interleukin-6, a transcription factor crucial for the regulation of IL-6 gene in hematopoietic cells (30). As a myeloid-specific transcription factor, NF-M binding activity is present only in chicken myelomonocytic cells and not in embryonic, erythroid, and liver cells, even though these cells were previously treated with LPS for 10 h. To our surprise, band shift assays with different nuclear extract preparations from DU249 cells failed to detect protein complexes binding to the elements D and E, although it is known that liver cells express C/EBP isoforms. In contrast, the human NF-IL6 is normally not expressed but can be activated by LPS in various tissues (30). The elements D and E of the -6.1 kb lysozyme enhancer confer myeloid-specific expression of the lysozyme gene (14). Our study shows for the first time that NF-M is not only necessary for myeloid-specific expression of the lysozyme gene but is also a modulator of the lysozyme expression in response to external stimuli such as LPS.
One mechanism by which LPS treatment can induce transcription of several target genes involves activation of the NF-KB transcription factor (39, 40). Recently, some LPS-induced DNA binding proteins were identified in B cells. For example, an LPS-inducible protein for a class I1 gene in B cells is distinct from NF-KB. This protein binds to two sites in a regulatory region of the B cell surface antigen la Ai gene, one of which contains the &-like binding site (41). Another LPS responsive factor, the LR1, present in pre-B and B cell lines but absent in primary B cells, binds to sequences from S1, S3, and Sa switch regions, and the heavy chain enhancer. It is suggested that LR1 might function as a transcription factor and play a role in LPS induction of heavy chain gene expression (27).
The mechanism by which LPS activates macrophages is poorly understood. Even less is known about signaling pathways that mediate LPS induction of the lysozyme gene expression in macrophages. Several studies demonstrated that treatment with LPS stimulated protein phosphorylation mediated via protein kinases C and tyrosine protein kinases in macrophages (42)(43)(44). NF-M translocates from the cytoplasm to the nucleus after treatment of primary chicken macrophages with LPS or 12-0-tetradecanoylphorbol-13-acetate (34). Although NF-M can be phosphorylated (34), it is likely that LPS-induced NF-M binding activity is not dependent on phosphorylation of the protein complexes in nuclei, but it might be controlled by phosphorylation of other nuclear factors like IKB, the inhibitor of NF-KB, that are present in cytoplasm of HDll cells. This question and that of which signaling pathway participates in the LPS-induced lysozyme gene expression remain to be answered.