Lipopolysaccharide-activated Kinase, an Essential Component for the Induction of the Antimicrobial Peptide Genes inDrosophila melanogaster Cells*

Eukaryotic organisms use a similar Rel/NF-κB signaling cascade for the induction of innate immune genes. InDrosophila, lipopolysaccharide (LPS) signal-induced activation of the Rel/NF-κB family transcription factors is an essential step in the transcriptional activation of inducible antimicrobial peptide genes. However, the mechanism by which the LPS-induced signaling pathway proceeds remains largely unknown. Here we have cloned a novel Drosophila LPS-activated kinase (DLAK) that is structurally related to mammalian IκB kinases. DLAK is expressed and transiently activated in LPS-responsiveDrosophila cells following LPS stimulation. Furthermore, DLAK can interact with Cactus, a Drosophila IκB and phosphorylate recombinant Cactus, in vitro. Overexpression of dominant-negative mutant DLAK (DLAKK50A) blocks LPS-induced Cactus degradation. DLAK-bound Cactus can be degraded in a LPS signal-dependent fashion, whereas the DLAKK50A mutant-bound Cactus is completely resistant to degradation in the presence of LPS. The DLAKK50A mutant also inhibits nuclear κB binding activity and κB-dependent diptericin reporter gene activity in a dose-dependent manner, but the κB-dependent diptericin reporter gene activity can be rescued by overexpression of wild type DLAK. Moreover, mRNA analysis of various κB-dependent antimicrobial peptide genes shows that LPS inducibility of these genes is greatly impaired in cells overexpressing DLAKK50A. These results establish that DLAK is a novel LPS-activated kinase, which is an essential signaling component for the induction of antimicrobial peptide genes following LPS treatment in Drosophilacells.


Eukaryotic organisms use a similar Rel/NF-B signaling cascade for the induction of innate immune genes. In Drosophila, lipopolysaccharide (LPS) signal-induced activation of the Rel/NF-B family transcription factors is an essential step in the transcriptional activation of inducible antimicrobial peptide genes. However, the mechanism by which the LPS-induced signaling pathway proceeds remains largely unknown. Here we have cloned a novel Drosophila LPS-activated kinase (DLAK) that is structurally related to mammalian IB kinases. DLAK is expressed and transiently activated in LPSresponsive Drosophila cells following LPS stimulation. Furthermore, DLAK can interact with Cactus, a Drosophila IB and phosphorylate recombinant Cactus, in vitro. Overexpression of dominant-negative mutant DLAK (DLAK K50A ) blocks LPS-induced Cactus degradation. DLAK-bound Cactus can be degraded in a LPS
signal-dependent fashion, whereas the DLAK K50A mutant-bound Cactus is completely resistant to degradation in the presence of LPS. The DLAK K50A mutant also inhibits nuclear B binding activity and B-dependent diptericin reporter gene activity in a dose-dependent manner, but the B-dependent diptericin reporter gene activity can be rescued by overexpression of wild type DLAK. Moreover, mRNA analysis of various B-dependent antimicrobial peptide genes shows that LPS inducibility of these genes is greatly impaired in cells overexpressing DLAK K50A . These results establish that DLAK is a novel LPS-activated kinase, which is an essential signaling component for the induction of antimicrobial peptide genes following LPS treatment in Drosophila cells.
The innate immune system of insects is a primitive but efficient host defense system aimed at preventing microbial invasion (1)(2)(3)(4). In response to microbial infection or to the presence of microbial cell wall components (lipopolysaccharide (LPS), 1 ␤-1,3-glucan, and peptidoglycan), insects rapidly induce the de novo synthesis of a battery of antimicrobial peptides that act in concert to suppress bacterial and fungal propagation (1,3,4). Analysis of the regulatory sequences of various insect antimicrobial peptide genes revealed that most of them share multiple B sites (5)(6)(7)(8). The B sites of insect immune genes have striking similarity to the NF-B-binding sites found in many inducible inflammatory genes in mammals (9). It was further demonstrated in Drosophila that LPS signal-dependent activation of a Rel/NF-B factor was an essential step for de novo synthesis of antibacterial peptides, such as diptericin, cecropin, attacin, drosocin, and defensin and an antifungal peptide, drosomycin (5,8,10). At present, three Rel/NF-B factors are found in Drosophila as follows: 1) Dorsal, a key regulator of dorsoventral polarity (11,12), 2) Dorsal-related immune factor (Dif), a factor implicated in the LPS-induced expression of certain antimicrobial peptides (5), and 3) Relish, a novel Rel/NF-B factor that also contains an IB-like domain (7).
Upon dorsoventral signaling, Dorsal is activated via the Toll pathway and then initiates the expression of various B-containing developmental genes (13)(14)(15). However, Dorsal Ϫ/Ϫ flies were shown to maintain the ability to regulate fully B-containing antimicrobial peptide genes (16), suggesting that other Rel/NF-B factors such as Dif and/or Relish might be implicated in immune signaling. Recently, some signaling components of the dorsoventral Toll pathway, including Spä zle 3 Toll 3 Tube 3 Pelle 3 Dif/Cactus (Dif instead of Dorsal), were also shown to play a role in the expression of the drosomycin gene but not in the expression of other antibacterial peptide genes such as diptericin or drosocin (16 -18). In the case of drosomycin and defensin, the activation of Dif is thought to be regulated through Toll and/or 18-wheeler (a Toll homolog) receptor-mediated signaling pathways (16,18,19). Such Toll family receptor-mediated B-dependent immune gene expression has also been shown to be involved in signal transduction in humans and in plants in response to microbial infection (20 -22). In humans, Toll-like receptor 2 (TLR2) and 4 (TLR4) are involved in LPS-induced activation of NF-B, which is required for immune gene expression (23)(24)(25). These recent findings suggest that a wide range of organisms use a common Rel/NF-B activation cascade for rapid induction of innate immune genes following an LPS challenge. In the mammalian innate immune system, NF-B is activated by a protein kinase CHUK/IB kinase (IKK) via phosphorylation of the cytoplasmic NF-B inhibitor, IB (26 -29). Phosphorylated IB is then rapidly degraded, and the free NF-Bs translocate to the nucleus to trigger B-dependent gene expression (9). Despite the striking similarity between human and Drosophila Rel/NF-B activation cascades, no IKK homologue has yet been found in Drosophila Rel/NF-B signaling. Recently, a powerful genetic approach was employed to identify novel genes involved in the activation of Rel/NF-B in Drosophila innate immune system. A number of immune response deficiency (ird) loci, which are essential for the nuclear import of Rel/NF-B and for the induction of diptericin reporter gene following bacterial challenge, have been identified by genetic screening of mutant flies (8). However, at present, the exact molecular identities of these loci and their mode of action on Rel/NF-B activation remain unknown. In the present study, we have isolated and characterized a novel Drosophila signaling mediator, DLAK, structurally and functionally related to the IKK family. DLAK is shown to be directly involved in LPS signal-dependent activation of Rel/NF-B and subsequent antimicrobial gene induction.

EXPERIMENTAL PROCEDURES
Drosophila Cell Culture-Drosophila immune-responsive Schneider cell line (ATCC CRL-1963) was grown at 23°C in Schneider medium (Sigma) supplemented with 10% fetal bovine serum (Life Technologies, Inc.), 50 units/ml penicillin, 50 g/ml streptomycin, and 50 g/ml gentamycin. Stably transformed cells expressing DLAK were maintained in the presence of 300 g/ml hygromycin.
cDNA Cloning-Except when specifically mentioned, all DNA and RNA manipulations were carried out using standard techniques (30). One Drosophila expressed sequence tag (EST) clone was identified from the Drosophila data base (BDGF Drosophila Genome Project, Berkeley, CA) and showed homology to amino acid sequences of human IKKs. With the primer pairs (sense, 5Ј-AGT GCC GAT CAG CAA GTC-3Ј, and antisense, 5Ј-CTC ATA GGC AAT TAC TCC G-3Ј) derived from the partial nucleotide sequence of the EST clone (LD09214), we amplified and isolated a specific 530-bp PCR product from an immune-stimulated Drosophila l (2)mbn cDNA library. The PCR fragment was labeled with 32 P by random priming and used as a probe to screen a Zap II l (2)mbn cDNA library. One positive clone was obtained from the Drosophila library after screening 2 ϫ 10 6 phages. A positive plaque was purified by two rounds of screening, and the cDNA insert was subcloned into pBlueScript SK(Ϫ) plasmid by in vivo excision as described by Stratagene. DNA sequencing of both strands was performed using an ALF-express™ automatic sequencer (Amersham Pharmacia Biotech).
In Situ Hybridization to Polytene Chromosomes-Polytene chromosomes from D. melanogaster (Oregon-R) were prepared from third instar larvae, as described by Pardue (38). Pre-hybridization treatment was performed as described with two modifications: the RNase treatment was omitted, and the polytene chromosomes were denatured by boiling for 3 min. The probes were a mixture of recombinant plasmids of DLAK cDNA and the white gene cDNA of Drosophila as a positive control. Probes were labeled with digoxigenin-dUTP (Roche Molecular Biochemicals labeling kit), as indicated in the manufacturer's instructions. The hybridization of the chromosomes on each slide was performed with 50 l of hybridization mixture (0.33 M NaCl, 5 mM MgCl 2 , 5 mM NaH 2 PO 4 , pH 7.2, 20 mM Tris-HCl, pH 6.7, containing 1ϫ Denhardt's solution) and 150 ng of each DNA probe. The hybridization mix was applied to the slide that was subsequently covered with a coverslip and sealed with Ross® rubber cement. The hybridization was carried out overnight at 65°C in a moist chamber. Post-hybridization treatment was as follows: 3 washings in 2ϫ SSC, 10 min each at 60°C, followed by 2 washes for 5 min in phosphate-buffered saline (PBS) and 1 wash for 2 min in PBS containing 0.1% Triton-X and 2 washings for 2 min in PBS. Subsequently, the hybridized probes were revealed essentially as described in the digoxigenin detection kit instructions (Roche Molecular Biochemicals). The anti-digoxigenin antibody conjugated to alkaline phosphatase was used at 1/2000 dilution in PBS containing 0.1% Tween 20. One hundred microliters was used for each slide and incubated for 1 h. The slides were washed four times in phosphate-buffered saline buffer containing 0.1% Tween 20 for 15 min and then twice in a solution of 100 mM NaCl, 50 mM MgCl 2 , 100 mM Tris-HCl, pH 9.5, 0.1% Tween 20 for 5 min. The alkaline phosphatase staining reaction was carried out for 2 h. Finally, the slides were washed twice for 15 min in TE buffer (10 mM Tris-HCl, pH 8.0, and 1 mM EDTA) and air-dried. The chromosomes were finally stained with 5% Giemsa solution.
Overexpression of DLAK and DLAK K50A in Drosophila Cells-In order to facilitate the detection, the open reading frame (ORF) corresponding to the full-length DLAK was hexahistidine-tagged in all experiments. This histidine-tagged DLAK ORF was subcloned into pMT/V5 vector (pMT/V5-His-DLAK) under the control of metallothionein promoter (Invitrogen). The catalytically inactive His-DLAK mutant construct (pMT/V5-His-DLAK K50A ) which replaced Lys 50 with Ala was generated by polymerase chain reaction-based mutagenesis (39). The sequence of all plasmids was confirmed by automated sequencing. Cell lines stably expressing His-DLAK or His-DLAK K50A were generated by transfection with pMT/V5-His-DLAK or His-DLAK K50A together with pCoHYGRO containing Escherichia coli hygromycin B-phosphotransferase gene under control of the Drosophila copia promoter (Invitrogen). Transfection was performed according to standard CaPO 4 protocols (40), and transfected cells were selected with hygromycin (300 g/ml) for 6 weeks. Expression was induced in cells by addition of CuSO 4 to the culture medium at a final concentration of 500 M. Various concentrations of CuSO 4 were used for electromobility gel shift assay. Cells were induced for 36 h before use.
Kinase Assay-Cell lines stably expressing His-DLAK or His-DLAK K50A were treated with 500 M CuSO 4 in order to initiate the production of His-DLAK or His-DLAK K50A for 36 h. Cell lines were then incubated in the presence of 10 g/ml LPS (Sigma; E. coli 026:B6) for 0, 5, 15, and 30 min. The cells were washed twice in Tris-buffered saline and solubilized in urea lysis buffer (8 M urea, 100 mM NaH 2 PO 4 , 10 mM Tris-HCl, pH 8.0) at room temperature for 30 min and subsequently centrifuged at 15,000 ϫ g for 15 min. The epitope-tagged DLAKs were purified by incubation with Ni ϩ -NTA-agarose resin for 1 h at room temperature. The resin was washed three times with washing buffer (8 M urea, 100 mM NaH 2 PO 4 , 10 mM Tris-HCl, pH 6.3) and finally eluted in SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer. For the in-gel kinase assay (41), affinity purified His-DLAK was resolved on a SDS-polyacrylamide gel. After electrophoresis, the gel was washed twice for 30 min at room temperature with 20% isopropyl alcohol, 50 mM Tris-HCl, pH 7.5, then twice for 30 min with buffer A (50 mM Tris-HCl, pH 7.5, 5 mM 2-mercaptoethanol). The gel was then incubated in 6 M urea in buffer A at room temperature for two 30-min incubations, followed by serial incubations in buffer A containing 0.05% Tween 20 and 3, 1.5, and 0.75 M urea, each for 30 min at room temperature. The gel was finally incubated in buffer A containing 0.05% Tween 20 at 4°C overnight for renaturation. After renaturation, the gel was incubated in kinase assay buffer (25 mM HEPES, pH 7.4, 25 mM ␤-glycerophosphate, 25 mM MgCl 2 , 2 mM dithiothreitol, 100 M ATP) containing 10 Ci of [␥-32 P]ATP per ml for 1 h at room temperature followed by extensive washing with excess amount of 5% trichloroacetic acid, 1% sodium pyrophosphate. The gel was then dried and subjected to autoradiography.
To perform the immunocomplex kinase assay, recombinant glutathione S-transferase (GST)-Cactus was expressed in E. coli and purified by using glutathione-Sepharose affinity chromatography according to the vendor's directions (Amersham Pharmacia Biotech). Purified GST-Cactus (5 g) was used as a substrate for the DLAK assay. Cells stably expressing His-DLAK were treated with 10 g/ml LPS. After treatment, His-DLAK from total cell extracts (100 g) was immunoprecipitated with monoclonal anti-His antibody, and the immunoprecipitates were subjected to an immunocomplex kinase assay essentially as described by Han et al. (42).
Reporter Gene Assay-His-DLAK K50A was subcloned into pPacPL vector (10) that contained the actin 5C promoter to generate pPacPL-His-DLAK K50A . The Dipt-B reporter construct contains eight copies of the B motif derived from the diptericin gene promoter, located upstream from the luciferase gene (6). LPS-responsive Schneider cells were cotransfected with 1 g of Dipt-B reporter construct together with different amounts (0, 1, 6, or 12 g) of pPacPL-His-DLAK K50A , 1 g of transfection efficiency vector pPacPL-␤-galactosidase and carrier DNA to the final concentration of 20 g. For the rescue experiment, cells were cotransfected with a constant amount (2 g) of pPacPL-His-DLAK K50A together with increasing amounts (0, 2, 6, or 12 g) of pPacPL-His-DLAK, 1 g of Dipt-B reporter DNA, 1 g of pPacPL-␤galactosidase and carrier DNA to the final concentration of 20 g. At 48 h after transfection, the cells were incubated with LPS (10 g/ml) for 7 h. Luciferase activity was measured according to the manufacturer's instructions (Promega). Luciferase activity was normalized to ␤-galactosidase activity to correct for variations in transfection efficiency.
Electromobility Gel Shift Assay-Cells stably transfected with pMT/ V5-His-DLAK K50A were treated with increasing concentrations of CuSO 4 (0, 100, 200, 500, and 1000 M) for 36 h to produce different amounts of His-DLAK K50A . Cells were then stimulated by addition of LPS (10 g/ml) for 1 h. Nuclear extracts were prepared from uninduced and LPS-induced cells as described previously (43). A synthetic oligonucleotide containing diptericin B sequence 5Ј-ATCGGGGATTC-CTTTT-3Ј was end-labeled by using T4 kinase in the presence of [␥-32 P]ATP (44). The binding reaction was performed for 30 min at room temperature by mixing 1 ng of purified 32 P-labeled diptericin B probe, 10 g of nuclear extracts, and 300 ng of poly(dI-dC) in 25 mM HEPES, pH 7.9, 150 mM KCl, 10% glycerol, 5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, as well as other protease inhibitors (Complete; Roche Molecular Biochemicals). Electrophoresis was performed under native polyacrylamide gel (6%) in 22.5 mM Tris-HCl, pH 8.0, 22.5 mM boric acid, 0.5 mM EDTA. The gel was dried and subjected to autoradiography.
Western Blot Analysis-Total cell extracts or purified His-DLAK was subjected to SDS-PAGE and transferred onto polyvinylidene difluoride membrane (45). The monoclonal anti-His antibody (Qiagen) was incubated overnight at 4°C. Anti-His antibody was used at 1:2000 dilution in TBST (50 mM Tris-HCl, pH 7.5, 140 mM NaCl, 0.02% Tween 20) in the presence of 5% bovine serum albumin. To analyze total cellular Cactus degradation, CuSO 4 -induced (500 M for 36 h) cells stably expressing His-DLAK K50A or uninduced cells were treated with LPS (10 g/ml) for 0, 15, 30, 60, 90, and 120 min. In some experiments, cells were pretreated with 120 M cycloheximide for 30 min as described by Samakovlis et al. (46) before LPS stimulation. Total cell extracts (100 g) were prepared and subjected to Western blot analysis to detect endogenous Cactus. The monoclonal anti-Cactus antibody was described by Whalen and Steward (47) and used at 1:100 dilution. A monoclonal anti-␣tubulin antibody (Oncogene) was used at 1:3000 dilution as a loading control. In the analysis of DLAK-bound Cactus degradation, cycloheximide pretreated, CuSO 4 -induced (500 M for 36 h) cells stably expressing His-DLAK or His-DLAK K50A were treated with LPS (10 g/ml) for 0, 30, 60, 90, and 120 min. His-DLAK (or His-DLAK K50A ) was precipitated using Ni ϩ -NTA-agarose resin and subjected to Western blot analysis using anti-Cactus antibody as described above. Following primary antibody, the blot was subsequently washed and incubated for 1 h with goat anti-mouse IgG conjugated to horseradish peroxidase diluted at The alignment is optimized by introducing gaps using CLUSTAL W (1.7) program. Comparative sequence analysis showed the kinase domain of DLAK to be most homologous to human IKK␤ (27,28) with 34% identity and 57% similarity, followed by human IKK␣ (26) with 33% identity and 54% similarity. The nucleotide sequence of DLAK has been submitted to the GenBank TM data base under accession number AF140766. 1:2000. Detection was carried out using the ECL Western blot detection kit (Amersham Pharmacia Biotech).

RESULTS
cDNA Cloning, Expression, Genomic Organization, and Polytene Chromosome Localization of DLAK Gene-In order to identify putative kinases that might be implicated in LPS-dependent Rel/NF-B activation, EST data bases (BDGF Drosophila genome project, Berkeley, CA) were screened for sequences homologous to human IKKs. We identified a novel partial Drosophila cDNA showing significant homology to human IKK ␤ (27,28). Screening of a Drosophila l (2) mbn library resulted in the isolation of a full-length cDNA clone. Analysis of nucleotide sequence revealed an ORF encoding 731 amino acids with a calculated molecular mass of 84,268 daltons. Sequence comparison showed that the kinase domain of DLAK is most homologous to human IKK ␤ (34% identity and 57% similarity) (Fig.  1). In addition, secondary structure analysis using the Pôle Bio-Infomatique Lyonnais program predicted a putative coiledcoils domain (leucine zipper) ( 469 LTKEQARYEMLVSGIN-ERALSLEDEMMEN) and a putative helix-loop-helix motif ( 610 NDAEEFAKSRFKLYN-625 EGEARHLPKSI-636 DHM-HYLY-FKT) at positions equivalent to those of IKK␣ and IKK␤ (26 -29).
In order to examine temporal expression of DLAK in Drosophila, RT-PCR analysis was performed and revealed that DLAK is constitutively expressed in all developmental stages ( Fig. 2A). DLAK expression was also detected in LPS-responsive Drosophila cells ( Fig. 2A).
To determine genomic structure of DLAK, we performed genomic Southern blot analysis. Genomic DNA was digested with a set of restriction enzymes, and fragments encoding DLAK were detected using random primed DLAK cDNA (nucleotide numbers 49 -2253) as a probe (Fig. 2B, upper panel). Based on the restriction sites of DLAK cDNA sequence and genomic Southern hybridization data, the structure of the genomic region encompassing the DLAK gene was predicted (Fig. 2B, lower panel). The restriction digest patterns strongly suggest that the DLAK gene is encoded by a single copy of the gene. In the case of the BamHI digestion (Fig. 2B, upper panel,  lane B), two fragments (0.5 and 1.8 kb) exhibited the same sizes corresponding to the length of BamHI fragments (nucleotide numbers 324 -846 and 846 -2145) deduced from DLAK cDNA sequence. This result suggests that this region of the DLAK gene (nucleotide numbers 324 -2145) is not interrupted by an intron. However, we do not exclude the possible existence of small size intron(s) that cannot be detected in our Southern blot analysis. In the 5Ј region of DLAK gene, the fragments digested by either SacI/PstI or HindIII/PstI (nucleotide number: 49 -421 and 67-421, respectively) gave a signal corresponding to 0.8 kb in our Southern blot analysis (Fig. 2B,  The Drosophila DLAK gene was cytologically localized by in situ hybridization on the right arm of chromosome 3, at position 89B (Fig. 2C).
DLAK Is Rapidly Activated in Response to LPS Stimulation and Activated DLAK Shows Autokinase Activity-To investigate whether DLAK responds to LPS stimulus in the Drosophila LPS-responsive Schneider cell line, we first transfected pMT/V5-His-DLAK or pMT/V5-His-DLAK K50A in order to generate stable cell lines expressing hexahistidine-tagged wild type DLAK (His-DLAK) or hexahistidine-tagged inactive mutant DLAK (His-DLAK K50A ) in which the conserved Lys 50 was mutated to Ala. Cells were stimulated with LPS for 0, 5, 15, and 30 min, and His-DLAK was purified by Ni ϩ -NTA-agarose resin. In order to measure LPS signal-dependent DLAK activity, purified His-DLAK or His-DLAK K50A from the above LPSstimulated cells was subjected to an in-gel kinase assay to detect autokinase activity. No DLAK activity was detected in the absence of LPS stimulation (Fig. 3A, upper panel). However, DLAK activity was detected in response to LPS treatment within 5 min, reaching its maximal level after 15 min and then returned to its basal level by 30 min (Fig. 3A, upper panel). The amount of His-DLAK in each sample was controlled by immunoblotting with a monoclonal anti-His antibody (Fig. 3A, lower  panel). In contrast to His-DLAK, the inactive mutant His-DLAK K50A showed no kinase activity in response to LPS treatment although it was expressed in equivalent amounts in the cell lysate (Fig. 3B, upper and lower panel). This result demonstrates that LPS stimulation transiently induces DLAK activation in Drosophila LPS-responsive Schneider cells and that activated-DLAK can be autophosphorylated in vitro.
Interaction, Phosphorylation, and Degradation of Drosophila IB Cactus by DLAK-In order to elucidate the possible functional role of DLAK in the regulation of IB proteins, we investigated whether DLAK could interact with Cactus, a Drosophila IB. Cell lysates from Schneider cells expressing His-DLAK were precipitated with Ni ϩ -NTA-agarose resin and analyzed by immunoblot analysis using a monoclonal anti-Cactus antibody. This antibody can specifically recognize a monospecific 71-kDa Cactus band (Fig. 4A). Fig. 4B shows that endogenous Cactus was coprecipitated with His-DLAK, whereas no Cactus band was observed in the absence of His-DLAK in mock-transfected, CuSO 4 -treated control Schneider cell lysates. This result indicates that Cactus binds DLAK at least when DLAK is overexpressed in Drosophila cells. However, interaction between DLAK and Cactus under normal in vivo conditions remains to be elucidated.
We then tested if Cactus could act as a substrate for DLAK by performing immune complex kinase assay. Schneider cells expressing His-DLAK were treated with LPS for 15 min, and activated DLAK was immunoprecipitated, and the kinase assay was performed using recombinant GST-Cactus. Fig. 4C shows that Cactus was phosphorylated by LPS-activated DLAK but not by unactivated DLAK. Control experiment showed that GST alone cannot be phosphorylated by LPSactivated DLAK (data not shown). These results indicate that the role of DLAK in LPS signaling may be achieved through its interaction with Cactus.
A previous study demonstrated that Cactus is normally expressed in the cytoplasm of immune cells and then rapidly degraded and resynthesized after immune challenge (17). Immune signal-induced degradation of Cactus is known to be essential for the activation of Rel/NF-B (17). Therefore we tested whether DLAK K50A could attenuate Cactus degradation induced by LPS. We examined the level of Cactus in the cells during LPS signaling in the presence or absence of DLAK K50A overexpression controlled by CuSO 4 . In the absence of DLAK K50A overexpression, the level of Cactus decreased at 15-30 min following LPS treatment but began to increase afterward (Fig. 4D, ϪCuSO 4 /ϪCHX). However, in the presence of DLAK K50A overexpression, no LPS-induced Cactus degradation was observed (Fig. 4D, ϩCuSO 4 /ϪCHX). The level of Cactus in LPS-treated DLAK K50A -overexpressed cells was even higher than untreated cells which might be explained by de novo synthesis of Cactus provoked by immune challenge (17). We also examined LPS-induced Cactus degradation following cycloheximide pretreatment in order to block de novo synthesis of Cactus during LPS treatment. Following cycloheximide pretreatment, the level of Cactus in control cells completely disappeared after LPS treatment (Fig. 4D, ϪCuSO 4 /ϩCHX). However, under the same conditions, LPS-induced Cactus degradation in DLAK K50A -overexpressed cells was much more attenuated (Fig. 4D, ϩCuSO 4 /ϩCHX). This result suggests that overexpression of DLAK K50A blocks LPS-induced Cactus degradation.
In order to confirm whether full DLAK activity is necessary for LPS-induced Cactus degradation, we examined the level of DLAK (or DLAK K50A -)-bound Cactus in the cells expressing His-DLAK (or His-DLAK K50A ) during LPS signaling in the presence of cycloheximide. Following LPS treatment, His-DLAK-bound Cactus was coprecipitated with His-DLAK using Ni ϩ -NTA-agarose resin and analyzed by Western blot analysis. The results show that the amount of His-DLAK-bound Cactus decreases indicating that LPS-induced Cactus degradation occurs (Fig. 4E). However, the amount of His-DLAK K50A -bound Cactus is nearly invariant (Fig. 4F) suggesting that LPS-induced Cactus degradation requires full DLAK activity. All these results suggest that DLAK is an essential signaling component in transmitting the LPS signal leading to Cactus degradation for Rel/NF-B activation.

DLAK K50A Blocks LPS-induced NF-B-dependent diptericin Reporter Gene Activity That Can Be Rescued by Overexpression of Wild Type DLAK-To investigate whether LPS-induced
DLAK activation is necessary for B transactivation, we performed a cotransfection assay. We used a luciferase reporter construct (Dipt-B-luciferase) containing eight copies of the 17-bp B-like sequence derived from the proximal upstream region of the LPS-inducible antibacterial diptericin gene (6). We analyzed the expression of luciferase from Dipt-B-luciferase constructs in the presence of a catalytically inactive DLAK as a dominant-negative regulator. For transient expression, His-DLAK K50A was subcloned into the Drosophila expression vector pPacPL under control of actin 5C promoter. In mock transfection using pPacPL with Dipt-B-luciferase, we generally obtained a 3-fold increase in luciferase activity following LPS stimulation in LPS-sensitive Schneider cells (Fig. 5A, upper panel). When increasing amounts of pPacPL-His-DLAK K50A were cotransfected with Dipt-B-luciferase, the level of reporter gene activity was inversely proportional to the amount of His-DLAK K50A (Fig. 5A, upper panel). The amount of His-DLAK K50A expression in each transfection experiment was verified by immunoblotting with a monoclonal anti-His anti-body (Fig. 5A, lower panel). High expression of His-DLAK K50A using 12 g of plasmid completely blocked LPS-induced diptericin reporter gene activity trans-activated by a Rel/NF-B family transcription factor. To investigate whether this dominant-negative effect is due to specific inhibition of the wild type DLAK function, we tested if Dipt-B-luciferase activity in DLAK K50A -overexpressed cells could be rescued by introducing wild type DLAK. Constant amount (2 g) of pPacPL-His-DLAK K50A was cotransfected with increasing amounts (0, 2, 6, and 12 g) of pPacPL-His-DLAK together with Dipt-B reporter gene. The result showed that the level of reporter gene activity was gradually rescued in proportion to the amount of His-DLAK (Fig. 5B). This result indicates that LPS-induced Dipt-B reporter expression can be specifically blocked by dominant-negative DLAK, and DLAK activity is seemingly necessary for LPS inducibility of Dipt-B reporter gene.
Inhibition of DLAK Activity Blocks LPS-induced Nuclear B Binding Activity-To investigate the role of DLAK activation in nuclear B binding activity, we performed an electromobility gel shift assay (EMSA) using the 17-bp B-like probe derived from the diptericin gene promoter. Previous studies have shown that this motif is homologous to the binding site for the mammalian NF-B and is essential for induction of the diptericin gene following LPS challenge (6). In order to block the DLAK pathway, we used an LPS-responsive Schneider cell line stably expressing dominant-negative DLAK under the control of the metallothionein promoter. Cells were induced for 36 h in the presence of increasing concentrations of CuSO 4 for the production of His-DLAK K50A . Cells were then stimulated for Lysates were immunoprecipitated using monoclonal anti-His-antibody, and kinase assay was performed using recombinant GST-Cactus as described under "Experimental Procedures." Reactions were stopped by adding SDS-PAGE sample buffer, and the samples were separated by electrophoresis and autoradiographed. D, dominant-negative DLAK inhibits LPS-induced Cactus degradation. Stable cell line expressing His-DLAK K50A under control of metallothionein promoter was treated with or without CuSO 4 (500 M) for 36 h. Cells were then incubated with LPS (10 g/ml) for 0, 15, 30, 60, 90, and 120 min. The level of Cactus was measured by performing immunoblot analysis using monoclonal anti-Cactus antibody. In the case of cycloheximide (CHX) pretreatment, cells were treated with 120 M cycloheximide for 30 min before LPS treatment. The same blot was also probed with monoclonal anti-␣-tubulin antibody in order to check the loading amounts. E, LPS-induced degradation of DLAK-bound Cactus. Lysates (100 g) from Schneider cells stably expressing His-DLAK (Sn-His-DLAK) were treated with LPS (10 g/ml) for 0, 30, 60, 90, and 120 min. Cells were treated with 120 M cycloheximide for 30 min before LPS treatment. Cactus was coprecipitated with His-DLAK using Ni ϩ -NTA-agarose. The precipitates were washed three times with lysis buffer, and the level of His-DLAK-bound Cactus was examined in an immunoblot analysis using monoclonal anti-Cactus antibody. The amount of His-DLAK was verified by an immunoblot analysis using monoclonal anti-His antibody. F, inhibition of LPS-induced degradation of DLAK K50A -bound Cactus. The same experiment was performed as described in E using His-DLAK K50Aoverexpressed cells (Sn-His-DLAK K50A ).
1 h by addition of LPS, and nuclear extracts were prepared for EMSA. We confirmed that the amount of His-DLAK K50A expression was proportional to the concentration of CuSO 4 by immunoblot analysis with a monoclonal anti-His antibody (Fig.  6, lower panel). EMSA data showed that nuclear B binding activity in response to LPS stimulation gradually decreased in a DLAK K50A expression-dependent manner (Fig. 6, upper and  lower panel). This correlation indicates that LPS-induced diptericin expression involves a B binding activity via the DLAK pathway. These results also prove that DLAK is an essential signaling component which leads to nuclear B binding activity in response to LPS.
LPS Inducibility of Various Antimicrobial Peptide Genes Is Impaired in DLAK K50A -overexpressed Cells-Previous studies provided evidence that the expression of various B-dependent Drosophila antimicrobial genes can be regulated independently by the Toll receptor-dependent pathway and/or the immune deficiency (imd) gene product-dependent pathway (8,16,18).
Antibacterial diptericin gene appears to be regulated mainly by the imd gene product, whereas antifungal drosomycin gene appears to be controlled by the Toll pathway. Other antibacterial peptide genes, such as cecropin, defensin, and attacin, require both the Toll pathway and imd gene product for their immune signal-induced expression. Hence, we examined whether the LPS inducibility of these antimicrobial genes is affected in the absence of DLAK activity using DLAK K50Aoverexpressed cells. We used semi-quantitative RT-PCR analysis using specific primers for each gene following LPS challenge. Fig. 7 shows that, in the absence of DLAK K50A expression, all examined antimicrobial peptide genes are upregulated 2.5-10 times (2.5 times for drosomycin expression, 4.6 times for diptericin expression, 6 times for defensin expression, 7.3 times for attacin expression, and 10 times for cecropin A expression) following LPS treatment. However, when DLAK K50A is overexpressed, LPS inducibility of diptericin, as well as cecropin, defensin, attacin, and drosomycin, is severely impaired (Fig. 7, A and B). defensin and attacin transcripts reach only 20% of control levels following infection, diptericin and cecropin transcripts reach 30% of control levels, and drosomycin transcripts reach 35% of control level (Fig. 7B). In control experiments, CuSO 4 treatment in untransfected control cells or a cell line stably expressing an unrelated Drosophila protein had no noticeable effect on the LPS inducibility of antimicrobial peptide genes (data not shown). This result suggests that both Toll-and imd-dependent pathways require DLAK activity for full LPS inducibility of various antimicrobial peptide genes. DISCUSSION A decade ago, the study of insect immune peptides and their genes was considered by most as an interesting novelty to understand better a primitive innate immune system and with hopes of generating possibly new therapeutic agents (48). During the same period, the study of Drosophila development was the focus of intense main stream research (49). At the time, no one could imagine the functional parallelism between developmental and immune processes in Drosophila. Then, although seemingly disparate, there appeared a striking similarity between Rel family transcription factors (Dorsal and NF-B) involved in the biological processes of dorsoventral pattern formation and immune response. This seeming disparity was bridged after NF-B-like molecules were proven to regulate the induction of antimicrobial peptide genes in Drosophila (5,6,50,51). This provided the impetus for in-depth biochemical, molecular, and genetic investigations into the relationship between the two systems (4 -8, 10, 16, 18, 19). These studies provided the experimental proof of the incredible but flagrant co-optation of the Toll pathway, initially governing the dorsoventral polarity in the Drosophila embryo, by the immune system of the insect, at least for the induction of certain antimicrobial peptides. Now, after a recent study by Meng et al. (18), there appears to be a cross-regulatory network involved in induction of multiple immune responsive genes.
In this context, we consider that our present results are very pertinent because no signaling kinase has yet been found for LPS-induced Rel/NF-B activation in Drosophila. During dorsoventral pattern formation, pelle kinase is known to be essential for developmental signal-dependent Dorsal activation via Cactus degradation (52,53). However, in LPS signal transduction, pelle is not required for the induction of diptericin (8,16,18), suggesting the existence of a kinase other than pelle might be implicated in LPS signal-dependent Rel/NF-B activation responsible for the induction of certain antibacterial peptides such as diptericin. Our results demonstrate that overexpression of a dominant-negative DLAK fully blocks LPS-induced Dipt-B reporter gene activity and also diminishes nuclear B binding activity. Furthermore, in addition to diptericin, other antimicrobial peptide genes such as cecropin, attacin, defensin, and drosomycin also lost their LPS inducibility in dominantnegative DLAK-overexpressed cells. DLAK appears to be a good candidate for this LPS signaling process. To date, only three Rel/NF-B family transcription factors, Dorsal, Dif, and Relish are known to be responsible for the induction of B motif-containing antimicrobial genes (5,7,11,12). We do not know exactly which of these transcription factors is responsible for each antimicrobial peptide. Very recently, in the contrast to the induction of drosomycin, the induction of diptericin, cecropin, and attacin genes was shown not to be affected in Dif and Dorsal double mutant flies (18). Involvement of a yet unknown Rel/NF-B transcription factor or Relish may participate in this process. With respect to the findings of Meng and colleagues (18), the exact molecular identities of the Rel/NF-B factors and its cytoplasmic inhibitor proteins with IB activity involved in the regulation of each antimicrobial gene remain to be determined. We do not yet know how DLAK intervenes in the activation of the putative Rel/NF-B, responsible for LPSinduced B-responsive antimicrobial gene induction. In the case of mammalian cells, the activation of Rel/NF-B transcription factors are thought to be regulated through the transient degradation of its cytoplasmic inhibitor protein (IB), where signal-induced phosphorylation of IB by specific IKKs initiates the inhibitor's conjugation to ubiquitin and subsequent degradation by the proteasome (54 -56). Given that DLAK has the most extensive sequence similarity with IKK and that DLAK exists in the form of a complex with Cactus and LPSinduced Cactus degradation is specifically blocked by dominant-negative DLAK, we could speculate that the dominantnegative form of DLAK fails to transmit the Cactus degradation signal, essential for Rel/NF-B activation and subsequently prevents LPS inducibility of at least certain B-dependent antimicrobial peptide genes. The possible effect of DLAK activity on the stability of other proteins with IB activity such as Relish is presently under investigation.
The essential involvement of DLAK for the induction of the B-dependent antimicrobial peptide genes is yet another example of the striking similarities between the LPS-induced signal transduction in Drosophila and mammalian IKK-mediated innate immune signaling reviewed by O'Neill and Greene (20). In mammals, two kinases responsible for NF-B activation (via IB phosphorylation) have been identified, IKK␣ and IKK␤ (26 -29). Both IKK␣ and IKK␤ contain three important functional domains as follows: a kinase domain for activity; a leucine zipper domain, necessary for dimerization; and a helix- FIG. 7. LPS-induced up-regulation of various B-dependent antimicrobial peptide genes is inhibited by His-DLAK K50A overexpression. A, LPS-induced expression of various B-dependent antimicrobial peptide genes in the presence or absence of DLAK K50A . Cells stably expressing His-DLAK K50A under control of metallothionein promoter were treated with or without CuSO 4 (500 M) for 36 h. Cells were then incubated with or without LPS (10 g/ml) for 6 h. The expression level of various antimicrobial peptide genes was measured by performing RT-PCR analysis as described under "Experimental Procedures." PCR products for each antimicrobial peptide gene (diptericin, cecropin, defensin, attacin, and drosomycin) and for ␤-actin as control gene were analyzed by agarose gel electrophoresis. The amount of expressed DLAK K50A was verified by immunoblot (IB) analysis using a monoclonal anti-His antibody. B, quantification of LPS-induced up-regulation of various B-dependent antimicrobial peptide genes in the presence or absence of DLAK K50A overexpression. The signals obtained from RT-PCR analysis were quantified by densitometer. Signal for each antimicrobial gene was normalized with the corresponding value of the ␤-actin signal. In each antimicrobial peptide gene, the expression level following LPS treatment in the absence of DLAK K50A was taken arbitrarily as 100, and the results are presented as relative expressions. Each bar represents the average of two independent experiments. loop-helix domain serving as a putative endogenous activator of IKK (26 -29). In the case of DLAK, its kinase domain is most homologous with that of IKK␤ (34% identity and 57% similarity). Furthermore, the analysis of the predicted secondary structure shows that DLAK also contains the putative structural domains (leucine zipper and helix-loop-helix) at equivalent positions to those in the mammalian homologs. In addition to these structural motifs, it was recently demonstrated in the case of IKK␤ that a serine cluster located between the helixloop-helix motif and the COOH terminus serves as a negative regulator of the kinase activity (57). We also detected a serine cluster (residues 673-705) containing 10 serine residues in the COOH terminus of DLAK corresponding to that of IKK␤. Very recent studies using IKK␣ Ϫ/Ϫ and IKK␤ Ϫ/Ϫ mice further confirmed the importance of the structural motifs and further identified the roles of IKK␣ and IKK␤ in development and inflammation signal transduction, respectively (57)(58)(59)(60). The strong structural resemblance between the mammalian IKK␤ and DLAK further corroborates our functional results indicating that DLAK is a Drosophila functional homolog of IKK␤. At present, it is unknown whether DLAK is also implicated in the regulation of B-responsive genes involved in early Drosophila development. In this study, DLAK was localized by in situ hybridization on chromosome 3R at position 89B. Further studies using DLAK Ϫ/Ϫ flies will elucidate the exact role(s) of DLAK in the Drosophila immune response and possibly in development.