[Truncated abstract] Osteoclasts are multinucleated giant cells responsible for the resorption of the mineralized bone matrix during the process of bone remodelling. During activation towards bone resorption, polarization of the osteoclast results in the formation of a unique plasma membrane, the ruffled border, the actual resorptive organelle of the osteoclast. Through this domain protons are actively pumped into the resorption lacuna creating an acidic microenvironment that favours the dissolution of the mineralized bone matrix. The polarised secretion of protons is carried out by the action of the vacuolar-type (H+)-ATPase (V-ATPase), composed of functionally and structurally distinct subunits of the V1 and V0 domains. The general structure of the V-ATPase complex is highly conserved from yeast to mammals, however, multiple isoforms for specific V-ATPase subunits do exist exhibiting differential subcellular, cellular and tissue-specific localizations. This study focuses on the molecular identification and characterization of V-ATPase accessory subunit Ac45 and the d2 isoform of the V0 domain d subunit in osteoclasts. Using the techniques of cDNA Subtractive Hybridization and DNA Micro-Array analyses respectively, the accessory subunit Ac45 and the d2 isoform of the V0 domain d subunit were identified in RAW264.7-cells derived OcLs. ... Using web-based computational predictions, two possible transmembrane domains, an N-terminus 'signal anchor' sequence and a C-terminus dilysine- like endoplasmic reticulum (ER) retention signal were identified. By confocal microscopy, EYFP-tagged e was found to localize to the perinuclear region of transfected COS-7 cells in compartments representing the ER and Golgi apparatus with some localization in late endosomal/lysosomal-like vesicles. ER truncation of e did not alter its subcellular localization but exhibited significantly weaker association with Ac45 compared to the wild-type as depicted by BRET analyses. Association with the other V0 subunits remain unaffected. This may hint at a possibility that Ac45 may play a role in the masking of the ER signal of e following it's incorporation into the V0 domain. Although no solid evidence for a role in the assembly of the mammalian VATPase have been established, subunit e still represents a potential candidate whose role in the V-ATPase complex requires further investigation. Collectively, the data presented in this thesis has provided further insight into the composition of the osteoclast V-ATPase proton pump by: 1) identifying an accessory subunit, Ac45 which shows promise as a potential candidate for the regulation and/or targeting of the V-ATPase complex in osteoclasts and truncation of its targeting signal impairs osteoclastic bone resorption; 2) identification and preliminary characterization of the d2 isoform of the V0 domain d subunit whose exact role in the V-ATPase complex and in osteoclasts remains to be determined, although its has been implicated to be essential for osteoclastic function; and 3) Preliminary characterization of subunit-e, a potential assembly factor candidate for the mammalian V-ATPase V0 domain.

Osteoclasts are multinucleated cells found exclusively in bone and are derived from the haematopoietic cells of monocytes/macrophage lineage. The cell-to-cell interaction between osteoblastic/stromal cells and osteoclast precursor cells is necessary for osteoclastogenesis. Receptor Activator of NF-κB ligand (RANKL) was identified as a membrane-bound TNF ligand family member that is the ‘master’ cytokine expressed on osteoblastic/stromal cells, which stimulate osteoclastogenesis through cell-to-cell contact with osteoclast precursors. RANKL is considered to be a factor that is necessary and sufficient for the induction of osteoclastogenesis (Lacey, et al., 1998). RANKL is a type II transmembrane cytokine of the TNF ligand superfamily and has an active TNF-like core domain at the extracellular domain. This active TNF-like core domain is thought to be the region through which it binds to it’s active receptor, RANK, for the activation of signal transduction pathways for the initiation of processes leading to osteoclastogenesis (Lacey, et al., 1998; Li, et al., 1999). It was hypothesized that any change in the active TNF-like core domain might affect the ability of RANKL binding to RANK and consequently affect the activation of signal transduction pathways and osteoclastogenesis. Hence, this thesis sought to investigate the effects of changes in the active TNF-like core domain by truncation mutation on the ability of RANKL binding to RANK and consequently affect the activation of signal transduction pathways and osteoclastogenesis. A cDNA fragment encoding the full-length TNF-like core domain of rat RANKL (rRANKL) (aa160-318) was cloned into the bacterial expression pGEX vectors and stably expressed in Eschechia coli as a fusion protein with the C-terminus of glutathione S-transferase (GST). Four mutants (aa160-302, aa160-268, aa239-318 and aa246-318) were also generated by truncation mutation in the TNF-like core domain, and cloned into the pGEX vector to produce GST-rRANKL mutants. The proteins were over-expressed and affinity purified to 95% in purity. GST-rRANKL (160-318) containing the full length TNF-like core domain was able to induced osteoclastogenesis in spleen cells in the presence of M-CSF and in RAW264.7 cells in the absence of M-CSF. It was also found to activate mature osteoclast activity in vitro, ex vivo and in vivo. It has the highest binding affinity to RANK and the greatest potency for NF-κB activation as well as the induction of osteoclastogenesis compared to the truncated mutants. Mutants generated by truncation of the TNF-like core domain revealed that the TNF-like core domain is important for the interaction with the RANK, for high binding affinity, NF-κB activation and induction of osteoclastogenesis. In general, the truncated mutants not only displayed a reduction in the binding affinity to RANK, but also a reduction in NF-κB activation, and significantly reduced potency in the induction of osteoclastogenesis. Interestingly, mutant GST-rRANKL (160-268) showed a higher affectivity than the other mutants did, in that it had greater binding affinity to RANK, and in NF-κB activation than the rest of the truncated mutants. Mutants GST-rRANKL (239-318) and GST-rRANKL (246-318) on the other hand, showed little potency in the induction of osteoclast formation, however, might have an inhibitory effect through competition with full length GST-rRANKL (160-318) as well as inducing a response in vivo resulting in an increase in the serum calcium level. In conclusion, this thesis demonstrated that the TNF-like core domain of RANKL is active, and imperative in the binding to RANK, activating signal transduction pathways and induction of osteoclastogenesis. Changes in the active TNF-like core domain affected the ability, affinity and efficiency of RANKL binding to the receptor, RANK and consequently affected the activation of signal transduction pathways and osteoclastogenesis.