rDRAK1, a Novel Kinase Related to Apoptosis, Is Strongly Expressed in Active Osteoclasts and Induces Apoptosis*

This is the first report of a novel serine/threonine kinase, rabbit death-associated protein (DAP) kinase-related apoptosis-inducing protein kinase 1 (rDRAK1), involved in osteoclast apoptosis. We searched for oste-oclast-specific genes from a cDNA library of highly enriched rabbit osteoclasts cultured on ivory. One of the cloned genes has a high homology with human DRAK1 (hDRAK1), which belongs to the DAP kinase subfamily of serine/threonine kinases. By screening a rabbit osteoclast cDNA library and 5 * -RACE (rapid amplification of cDNA ends), we obtained a full length of this cDNA, termed rDRAK1. The sequencing data indicated that rDRAK1 has 88.0, 44.6, 38.7, and 42.3% identity with hDRAK1, DAP kinase, DRP-1, and ZIP (zipper-interact-ing protein) kinase, respectively. To clarify the role of DRAK1 in osteoclasts, we examined the effect of three osteoclast survival factors (interleukin-1, macrophage colony-stimulating factor, and osteoclast differentia-tion-inducing factor) on rDRAK1 mRNA expression and the effect of rDRAK1 overexpression on osteoclast apoptosis. The results suggested that these three survival factors were proved to inhibit rDRAK1 expression in rabbit osteoclasts. After transfection of a rDRAK1 expression vector into cultured osteoclasts, overexpressed rDRAK1 was localized exclusively to the nuclei and induced apoptosis. Hence, rDRAK1 may play an important role in the core apoptosis program in osteoclast.

Osteoclasts are multinucleated giant cells with the resorbing activity of calcified tissues. Stem cells for osteoclasts are present in the hematopoietic tissue (1). How the mononuclear stem cells fuse and differentiate into mature osteoclasts has begun to be clarified with the identification of osteoclast differentiationinducing factor (ODF) 1 /receptor activator of nuclear factor-B (NF-B) ligand (RANKL)/tumor necrosis factor-related activation-induced cytokine (TRANCE)/osteoprotegerin ligand (OPGL) (2)(3)(4)(5), and osteoclastogenesis inhibitory factor (OCIF)/ osteoprotegerin (OPG)/tumor necrosis factor receptor-like molecule 1 (TR1) (6 -8). After maturation, osteoclasts undergo a multistep resorption process (resorption cycle) that involves matrix recognition, osteoclast attachment, polarization and formation of a sealing zone on the bone, and resorption itself, with final detachment and possible cell death (9,10). Apoptosis of osteoclasts is a hot topic experimentally and clinically in bone biology for the following reasons. It has been reported that osteoclasts die by apoptosis at the end of the bone resorption process (11) and that control of apoptosis might potentially represent a key step in the regulation of bone resorption (remodeling). It has been suggested that the drugs with the most potential in the treatment of osteoporosis, bisphosphonates (12) and estrogens (13), act as inducers of osteoclast apoptosis.
For several years, we have tried to clone osteoclast-specific genes involved in osteoclast resorption (cycle) including apoptosis. For this purpose, it was necessary to isolate primary osteoclasts of sufficient numbers and purity for biochemical or molecular biological study. Recently, an efficient method for the isolation of rabbit osteoclasts, established by Kakudo et al. (18), has contributed to many studies on osteoclasts (19 -24). Using a cDNA library of highly enriched rabbit osteoclasts, we obtained several fragments of genes that might be related to the osteoclast resorption cycle. One of them has high homology * This work was supported in part by Grant JSPS-RFTF96I00202 for the "Research for the Future Program" from the Japan Society for the Promotion of Science. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB042195 (rDRAK1). ‡ The first three authors contributed equally to this work. ¶ Domestic Research fellow, Japan Science and Technology Corporation (JST).
‡ ‡ To whom correspondence should be addressed. with hDRAK1, a member of the DAP kinase subfamily (25,26), and it was termed as rDRAK1. DAP kinase is a calcium/calmodulin-regulated serine/threonine protein kinase and an effector of ␥-interferon-mediated apoptosis in HeLa cells (26). To demonstrate the role of rDRAK1 in rabbit osteoclast apoptosis, we identified the full sequence of rDRAK1 and verified the effect of typical osteoclast survival factors (IL-1, M-CSF, and ODF) on DRAK1 mRNA expression by a quantitative reverse transcription polymerase chain reaction (RT-PCR) method. We also examined the overexpressed rDRAK1-inducing apoptosis in rabbit osteoclasts by terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end labeling (TUNEL).

EXPERIMENTAL PROCEDURES
Isolation of Osteoclasts-Osteoclasts were harvested from 10-day-old Japanese white rabbits according to the method of Kakudo et al. (18) with minor modifications. Briefly, the long bones were minced with scissors for 15 min in ␣-minimum essential medium (␣-MEM, Life Technologies, Inc.) supplemented with 5% fetal bovine serum (FBS, Life Technologies, Inc.). The unfractioned cells were collected and seeded on collagen gel (Nitta Gelatin Inc.)-coated dishes. After 3 h of incubation, the cells were treated with phosphate-buffered saline (PBS) containing 0.001% Pronase (Sigma) and 0.02% EDTA for 15 min at room temperature to remove the nonadherent hematopoietic cells; this was followed by treatment with 0.01% collagenase (Wako Pure Chemical Industries, Ltd.) for 5 min at room temperature to further remove most of the stromal cells. Finally, osteoclasts were obtained in suspension by treatment with 0.1% collagenase at 37°C for 10 min. The yield and purity of osteoclasts from one rabbit was 1-3 ϫ 10 4 and more than 90%, respectively. Purified osteoclasts were seeded on round ivory slices (6.5-mm radius) in four kinds of medium: ␣-MEM with 5% FBS supplemented with 1.0 ng/ml IL-1 (mouse recombinant, Life Technologies, Inc.), 50 ng/ml M-CSF (mouse recombinant, Calbiochem), or 500 ng/ml ODF (mouse recombinant, Calbiochem) and no supplement. When the cells were prepared for transfection, the use of 0.001% Pronase, 0.02% EDTA and 0.01% collagenase was omitted. After 3 h of incubation, the nonadherent cells were washed out with PBS, and the osteoclasts were then collected with 0.1% collagenase. This short procedure yielded 2-5 ϫ 10 4 osteoclasts/rabbit with 30 -50% purity.
Construction of an Osteoclast cDNA Library and cDNA Cloning-The purified rabbit osteoclasts were cultured on sliced ivory with ␣-MEM containing 5% FBS for 3 h at 37°C. Total RNA was then extracted and further purified with TRIzol ® reagents (Life Technologies, Inc.) according to the manufacturer's instructions. Following reverse transcription of the RNA, we constructed a cDNA library of highly enriched rabbit osteoclasts cultured on ivory by using THE SMART cDNA Library Construction Kit (CLONTECH). The cDNA fragment cloned into The pCR-TRAP vector (GenHunter Co.) was sequenced using an automated DNA sequencer (ABI model 310, PerkinElmer Applied Biosystems). Then, the full-length cDNA of the gene was obtained by screening the rabbit osteoclast cDNA library and the 5Ј-rapid amplification of cDNA ends (5Ј-RACE) according to the manufacturer's instructions (Life Technologies, Inc.). The cDNA obtained was labeled with a DIG DNA labeling kit (Roche Diagnostics) and used as probe for tissue distribution. The determined DNA sequences were subjected to a homology search against the BLAST nucleotide sequence data base.
Tissue Distribution Determined by Slot Blot Analysis-The total RNA of rabbit tissue (bone marrow, brain, lung, liver, and kidney) was isolated and further purified by the TRIzol method. A nylon membrane (Amersham Pharmacia Biotech) of suitable size was used after soaking in 10ϫ standard saline citrate (SSC) for 10 min. The RNA (10 g) in denatured solution was heated at 65°C for 15 min and then immediately chilled on ice. The RNA mixture was blotted onto nylon membrane followed by heating membrane at 80°C for 2 h. Hybridization with standard hybridization buffer containing digoxigenin (DIG)-labeled probe at a concentration of 2 g/ml was carried out overnight at 65°C. After washing the membrane, the color was developed with a DIG DNA detection kit (Roche Diagnostics). The obtained cDNA clone was prepared as DIGlabeled probe with a DIG DNA labeling kit. The density of the positive spots was analyzed with NIH Image software. In this case, DIG-labeled ␤-actin was used as an internal standard. Each experiment was done twice (two batches using two rabbits) and independently (n ϭ 2).
Quantitative RT-PCR-Total RNA was isolated from osteoclasts using TRIzol ® reagent. For reverse transcription, the reaction mixture contained 2 g of RNA, 2.5 M oligo(dT) primer, and 5 units of avian myeloblastosis virus reverse transcriptase (AMV, TaKaRa) in a total volume of 20 l. The reaction was performed by incubation for 1 h at 42°C and stopped by heating for 5 min at 99°C. Aliquots (0.5 l) of the reverse transcriptase products were amplified in the reaction mixture (20 l) containing LightCycler-FastStart DNA Master SYBR Green I, 0.5 M each primer, and 3 mM MgCl 2 using LightCycler (Roche Molecular Biochemicals). After pre-incubation at 95°C for 10 min, a PCR was performed with 40 cycles of denaturation at 95°C for 15 s, annealing at 60°C for 5 s, and elongation at 72°C for 10 s. A single fluorescence reading was taken in each cycle following the elongation step. The primers used were as follows: rDRAK1, 5Ј-CGTGGTTGACACAGAGC-AGT-3Ј (corresponding to 905-924) and 5Ј-TTCGGTTCCTGGTTTCTC-AG-3Ј (1016 -1035); rabbit glyceraldehyde-3-phosphate dehydrogenase as an internal standard, 5Ј-CGACATCAAGAAGGTGGTGA-3Ј and 5Ј-CCAGCATCGAAGGTAGAGGA-3Ј. The GenBank TM accession number for rabbit glyceraldehyde-3-phosphate dehydrogenase is L23961. The relative rDRAK1 mRNA levels for each condition were determined by performing quantitative RT-PCR three times for each of the six independent batches (n ϭ 6).
Transfection and Immunostaining-We subcloned the full-length rDRAK1 and LacZ into the pcDNA4/HisMaxA expression vector (Invitrogen) to create His-tagged rDRAK1 and LacZ expression vectors. The semi-purified osteoclasts were cultured on FBS-coated 4-well glass chamber slides (Nunc) for 24 h at 37°C. Transfection was done by using Tfx-50 reagent (Promega). After the cultivation period, the cells were incubated in the presence of 0.1-0.5 g of His-rDRAK1 or His-LacZ in ␣-MEM containing 5% FBS and 100 nM Tfx-50 and then cultured for 6 h at 37°C. After 6 h of incubation, the cells were fixed with 3.7% formaldehyde (TAAB Laboratories Ltd.) and were blocked with 5% skim milk containing 0.05% Triton X-100. They were then incubated with mouse anti-6-His monoclonal antibody (Babco Covance Co.) at 4°C overnight. After being washed with 0.1% Tween 20 in PBS three times, the cells were stained with goat anti-mouse fluorescein isothiocyanateconjugated secondary antibody (Jackson ImmunoResearch Laboratories) for 2 h at room temperature. After three more washes, the nuclei were stained with 10 g/ml propidium iodide (PI, Sigma), and the samples were embedded in 0.1% paraphenylenediamine (Sigma) in glycerol. The cells were then observed under a laser confocal scanning microscope (MRC-600 UV, Bio-Rad). After a reaction using nitroblue tetrazolium chloride/5-bromo-4chloro-3-indolyl-phosphate (NBT/BCIP, Roche Diagnostics) and 4-toluidine salt as substrates for alkaline phosphatase, the stained cells were analyzed under a light microscope.

RESULTS
Cloning of rDRAK1-We have cloned several gene fragments of osteoclast, which might be related to each resorption process (27). One of these fragments has a sequence of 291 base pairs located at the 3Ј-terminus as shown in Fig. 1. By comparison with the nucleotide data base of BLAST, it was found to have high homology with hDRAK1, in which the catalytic domain is related to that of DAP kinase, a serine/threonine kinase involved in apoptosis. In this paper, we focus on this gene and its role in osteoclast apoptosis. To obtain a full-length copy of the fragment related to hDRAK1, screening of a rabbit osteoclast cDNA library and 5Ј-RACE were performed. This cDNA library was constructed from osteoclasts attached to ivory and included 1.7 ϫ 10 6 independent clones. For 5Ј-RACE, the RNA extracted from bone marrow cells was used as a template. The predicted amino acid sequence of the obtained full-length cDNA, termed rDRAK1, is 397 residues long (Fig. 1) and is 88.0, 58.4, 44.6, 38.7, and 42.3% identical to hDRAK1, hDRAK2, DAP kinase, DRP-1, and ZIP kinase (25, 26, 28 -30), respectively. The putative kinase domain of rDRAK1 is located at the N terminus and contains all 11 subdomains conserved among serine/threonine kinases (Fig. 2).
Tissue Distribution of rDRAK1-Comparing rDRAK1 expression in six kinds of tissue (bone marrow, brain, heart, lung, liver, and kidney), the expression is highest in bone marrow as shown in Fig. 3. This result is consistent with the presence of mature osteoclasts in bone marrow tissue. DRAK1 expression levels in brain and lung are similar, and the expression in heart is the weakest.
Effect of Osteoclast Survival Factors on mRNA Expressions of rDRAK1-To elucidate the role of DRAK1 in osteoclast apoptosis, it is necessary to examine the effect of osteoclast survival factors on DRAK1 mRNA expression. However, to carry out such a series of experiments, the osteoclast mRNA from 10 rabbits was needed. The usual semi-quantitative RT-PCR FIG. 2. Alignment of amino acid sequences of rDRAK1, hDRAK1/2, DAP kinase, DRP-1, and ZIP kinase. The sequences of rDRAK1, hDRAK1/2, DAP kinase, DRP-1, and ZIP kinase with cDNA-deduced amino acid sequences are described (25,26,28,29,30). The 11 subdomains are conserved among these kinases. Identical amino acid residues are indicated by solid boxes. Gaps are indicated by hyphens. The GenBank TM accession numbers for rDRAK1, hDRAK1, hDRAK2, DAP kinase, DRP-1, and ZIP kinase in DDBJ/GenBank TM / EMBL are AB042195, AB011420, NM004226, NM004938, NM014326, and AB022341, respectively.

FIG. 3. Tissue distribution of rDRAK1 in rabbit.
The relative rDRAK1 mRNA levels in rabbit tissues (bone marrow, brain, heart, lung, liver, and kidney) are shown. The RNA of each tissue was hybridized on nylon membrane with a DIG-labeled DRAK1 probe and subjected to color development with a DIG DNA detection kit. A ␤-actin probe was used as an internal standard. Each experiment was done twice (two batches using two rabbits) independently (n ϭ 2). The data are given as the mean Ϯ S.E. rDRAK1 Induces Apoptosis in Osteoclasts method was not adequate. Therefore, we adopted a quantitative RT-PCR method for this analysis, as described under "Experimental Procedures," using a LightCycler system and reagents (Roche Molecular Biochemicals). Typical examples of the use of this system are cited in the references (31,32). We chose IL-1, M-CSF, and ODF as typical osteoclast survival factors. Fig. 4A shows the effect of the osteoclast survival factors on DRAK1 mRNA expression. The three factors inhibited DRAK1 expression by 51, 34, and 57% (IL-1, M-CSF, and ODF, respectively). We observed the inhibition of DRAK1 expression by survival factors to be partial. With survival factors, DRAK1 is still expressed in osteoclasts. At the same time, optical microscopic images are shown in Fig. 4B, suggesting that those three factors induced spreading of osteoclasts.
Cellular Localization of Ectopically Expressed rDRAK1 in Osteoclasts-It has been reported that hDRAKs are exclusively localized to the nuclei (25). To investigate the cellular localization of rDRAK1, the expression plasmids His-rDRAK1 and His-LacZ were transiently transfected into osteoclasts. After a cultivation period, the His-tagged proteins were detected by indirect immunostaining with anti-His monoclonal antibody and fluorescein isothiocyanate-conjugated secondary antibody. All nuclei could be simultaneously visualized by staining with propidium iodide. As seen in Fig. 5, a-c, the fluorescent signals showed that LacZ localized in the cytoplasm but not in the nucleus. On the other hand, in rDRAK1-transfected osteoclasts, the fluorescent signals and the nuclei stained by propidium iodide overlapped as shown in Fig. 5, d-f. These results indicated that rDRAK1 is exclusively localized to the nuclei.
Overexpression of rDRAK1 Induced Apoptosis in Osteoclasts-Most of the DAP kinase family has the ability to induce apoptosis in various cells (25,26,28,29,33). It was reported that DAP kinase participates in tumor necrosis factor-␣ and Fas-induced apoptosis (34). To investigate whether rDRAK1 induces apoptosis in osteoclasts, we performed a TUNEL assay to detect DNA fragmentation after transfection. As shown in Fig. 6, the nuclei of rDRAK1-transfected osteoclasts were stained strongly, indicating that the nucleus condensed and was then fragmentated. Chromatin condensation followed by fragmentation is a specific morphological characteristic of apoptosis; therefore, rDRAK1 induced apoptosis in osteoclasts when it was overexpressed. On the other hand, overexpression of His-LacZ did not affect the DNA fragmentation or the nuclear morphology of osteoclasts.

DISCUSSION
This is the first report of a novel serine/threonine kinase, rDRAK1, involved in osteoclast apoptosis. In the past 5 years, much work has been done relating to osteoclast apoptosis of the basic and clinical level, with many studies using bisphosphonates. Bisphosphonates are known to inhibit bone resorption and to be therapeutically effective in diseases involving increased bone turnover, such as Paget's disease. Apoptosis has  osteoclast morphology (B). Highly enriched osteoclasts were cultured on ivory slices for 3 h in the four kinds of medium: ␣-MEM with 5% FBS supplemented with 1.0 ng/ml IL-1, 50 ng/ml M-CSF, or 500 ng/ml ODF, and no supplement. Total RNA was isolated by the TRIzol method. The relative rDRAK1 mRNA levels for each condition were determined by performing quantitative RT-PCR three times for each of six independent batches (n ϭ 6) as described under "Experimental Procedures." A, rDRAK1 mRNA level; B, optical microscopic images of osteoclasts. Bar, 10 m. The data are shown as the mean value Ϯ S.E.

FIG. 5. Cellular localization of the ectopic expression of rDRAK1 on osteoclasts.
Osteoclasts were transfected with 0.5 g of pcDNA4/HisMaxA-rDRAK1 (His-rDRAK1) or LacZ (His-LacZ) plasmid, which contain His tag. Six hours post-transfection, the cells were immunostained using mouse anti-6-His monoclonal antibody followed by goat anti-mouse fluorescein isothiocyanate-conjugated secondary antibody. Micrographs show epitope staining (left) of LacZ (a-c) or rDRAK1 (d-f), propidium iodide staining (middle), and a composite of the two panels (right). Bar, 50 m. rDRAK1 Induces Apoptosis in Osteoclasts proved to be a major mechanism whereby bisphosphonates reduce osteoclast numbers and activity (12,35). Further study of bisphosphonate-induced signaling and its relation to the mevalonate pathway (36 -39), as well as caspase cleavage of MST1 kinase (mammalian STE20-like kinase 1) (38), is underway. Moreover, the action of bisphosphonates in bone metastasis has attracted much attention in cancer researchers (40,41). Suppression of osteoclast activity is a primary approach to inhibiting bone metastasis, and bisphosphonates, specific inhibitors of osteoclasts, have been used widely for treatment of bone metastasis in cancer patients. Despite the clinical and basic importance of bisphosphonates in osteoclast apoptosis, the molecular mechanism inducing apoptosis has not been fully elucidated, and no report has been published regarding the core apoptosis program in osteoclasts. In addition to bisphosphonates, estrogen (13,15), high extracellular calcium ions (42,43), and mechanical force (44) have been reported to induce osteoclast apoptosis.
IL-1 (45,46), M-CSF/colony stimulating-factor-1 (47), and ODF/RANKL/TRANCE/OPGL (48), by contrast, have been reported to support osteoclast survival by preventing apoptosis. Jimi et al. (45) found that IL-1 transiently activates transcription factor NF-B found in the osteoclast nuclei and involved in the survival of osteoclasts. M-CSF plays important roles throughout the life of the osteoclast including in proliferation, differentiation, migration, chemotaxis, and survival (49). Both IL-1 and M-CSF induce the multinucleation of pre-osteoclasts through their respective receptors; however, actin ring formation (a functional marker of osteoclasts) and pit-forming activity of multinucleated cells have been observed in pre-osteoclast culture treated with IL-1 but not with M-CSF (46). These results suggest that IL-1 induces the bone resorbing activity of osteoclasts but M-CSF does not. Okahashi et al. (50) reported the involvement of caspase in the regulation of osteoclast survival. They found that two peptide inhibitors of caspase extended the survival time of osteoclast-like cells and the effect was enhanced by co-addition with IL-1 or M-CSF. They suggest that caspase-3 is particularly important in osteoclast apoptosis. ODF targets osteoclast precursors and osteoclasts to enhance differentiation and activation. The effect of ODF on osteoclast survival was reported by Lacey et al. (48). According to their report, apoptotic change with elevated caspase-3 activity was inhibited by a combination of caspase-3 inhibitor and ODF. The signaling/survival pathways of both ODF and M-CSF are required for optimal osteoclast survival. They also reported that M-CSF maintained NF-B activation and increased the expression of bcl-2 but had no effect on c-Jun N-terminal kinase (JNK) kinase activation. In contrast, ODF enhanced both NF-B and JNK kinase activation and increased the expression of c-src but not bcl-2. Therefore ODF is essential but not sufficient for osteoclast survival, and endogenous M-CSF levels are insufficient to maintain osteoclast viability in the absence of ODF. These findings indicate that IL-1, M-CSF, and ODF support osteoclast survival but differ in their manner or mechanism. The results outlined in Fig. 4 show that the expression of DRAK1 was inhibited partially by each of the three survival factors, IL-1, M-CSF, and ODF. This finding would imply that DRAK1 contributes to the survival mechanism through NF-B activation or apoptosis induced by caspase-3 activation. The inhibitory effects on the mRNA expression of DRAK1 might therefore be a common way for maintaining osteoclast survival. DRAK1 and DRAK2 were first cloned by Sanjo et al. (25) in 1998 as novel serine/threonine kinases. Both DRAKs are composed of an N-terminal catalytic domain and a C-terminal domain responsible for the regulation of kinase activity. The kinase activity of DRAK1 is significantly stronger than that of DRAK2. How, specifically, is DRAK1 expressed in osteoclasts? Sanjo et al. (25) reported that overexpression of DRAK1 induced morphological changes in NIH3T3 cells that were associated with apoptosis but not in COS-7 cells, suggesting that the sensitivity to DRAK1-induced apoptosis differed depending on cell type. According to our results, DRAK1 is strongly expressed in bone marrow tissues. Moreover, DRAK1 is expressed weakly or not at all in osteoblasts (data not shown); however, it is expressed strongly in osteoclasts. These results suggest that DRAK1 is closely involved in the regulation of osteoclastogenesis and osteoclast apoptosis. How does DRAK1 act in osteoclast apoptosis? DRP-1 (28), DRAK1/2, and ZIP kinase (29) form a DAP kinase subfamily of serine/threonine kinases that mediate apoptosis. As shown in Fig. 2, all the members have a similar kinase domain at their N terminus. DAP kinase (26,34,51) and DRP-1 have a calcium/calmodulin regulatory region and are localized to the cytoplasm, whereas DRAK1/2 and ZIP kinase have no calmodulin region and are localized to the nuclei. Overexpression in the wild type and kinase inactive mutant suggests that their kinase activity triggers apoptosis. Among, DAP, DRP-1, DRAK1/2, and ZIP kinase, the noncatalytic C-terminal regions are structurally different and do not share any homology. DAP kinase contains two known domains characterized by eight ankyrin repeats and the death domain, whereas ZIP kinase has a leucine zipper domain at its C-terminal end. These domains are known to mediate protein-protein interaction. It is suggested that DAP kinase is activated by the formation of a homodimer or binding to regulatory molecule(s) and ZIP kinase is activated when it FIG. 6. rDRAK1-induced osteoclast apoptosis. Osteoclasts were transfected and stained by TUNEL assay as described under "Experimental Procedures." Micrographs show the results of the TUNEL assay to detect DNA fragmentation. Arrows show condensed chromatin. Overexpression of His-LacZ did not affect the morphology of osteoclasts (left), but that of His-rDRAK1 induced an apoptotic nuclear morphology (right).

rDRAK1 Induces Apoptosis in Osteoclasts
homodimerizes through the leucine zipper domain. The C-terminal end of DRP-1 is also essential for its dimerization, and the homodimerization is a requirement for the functionality of this kinase in apoptosis. By analogy with other DAP kinases, the noncatalytic C-terminal region of DRAK1 may function as an interaction domain important for induction of apoptosis through homodimerization or binding to some regulatory molecule(s). Our DRAK1 overexpression experiment suggests that DRAK1 was exclusively localized to the nuclei, consistent with the results of Sanjo et al. (25), and at the same time, apoptosis was clearly observed by TUNEL. This means that DRAK1 triggers osteoclast apoptosis in the nuclei by its kinase activity, perhaps through homodimerization or binding to specific regulatory molecules. According to our results, rDRAK1 is expressed in active osteoclasts, even with survival factors, which also suggests the existence of an inhibitory (regulatory) molecule for rDRAK1-induced apoptosis. However, the upstream cascade of the DAP kinase family including DRAK1 and their substrates are unknown. In general, serine/threonine kinases function as molecular gatekeepers guarding entry into the death pathway, and their substrates can either enhance or inhibit susceptibility to apoptosis (52). Moreover, no evidence has been presented that kinases are required for the execution of apoptosis. The caspase family appears to be the executioners of apoptosis, although none has been shown to be regulated by phosphorylation. It is likely that kinases play important roles in signaling cascades linking exogenous stimuli to the core apoptosis program. rDRAK1 might be a candidate for a key effective molecule of the apoptotic signaling cascade, and more experiments on DRAK1 might open the way for elucidating the core apoptosis program in osteoclasts.