Fig 1.
OC-STAMP knockout allele, radiographs.
A. Gene targeting vector is a “knockout first” homologous recombination gene trap with a selection cassette and recombination sites for Cre (loxP) and flippase (Frt), as indicated. This allele, which results in a complete loss-of-function, is designated Ocstamp tm1a(KOMP)Wtsi. Exons in the upper (gene) and lower (targeting construct) diagrams are numbered 1, 2, and 3. The selection cassette in intron 1 eliminates production of any functional OC-STAMP. The 5’ and 3’ UTR’s are shown in boxes, and the 3’ alternative end is shown in light blue. Asterisks indicate locations of genotyping primers. From left to right they are: OC-F1 (red) (5’-TTGCCTGTAAATGATGGAGTGGGC-3’); En2R1 (blue) (5’-TGGTGTGGGAAAGGGTTCGAAGTT-3’); OCR1E (purple) (5’-TGGCGCAGCTGGTAAGTGGTATTA-3’). These give PCR products of 1044 bp from WT (OCF1 and OCR1E) and 297 bp from the knockout allele (OCF1 and En2R1). To confirm correct 3’ end insertion, the right-hand pair of primers are: LoxPF (green) (5’-GAGATGGCGCAACGCAATTAAT-3’) and SR1 (brown) (5’-CTGTGACTAAGTAACCATCAAAGCGG-3’), which give a 681 bp product only in the targeted allele (not shown). B. PCR of genomic DNA from the mice yielded the expected sizes in homozygous wild type (+/+; upper arrowhead indicates 1044 bp), heterozygotes (+/-), and homozygous knockout (-/-; lower arrowhead indicates 297 bp). C. X-rays of 2-week-old wild type (left) and OCSt-KO mice showed no differences in either whole-body views (upper) or rear quadrant views (lower panels). All skeletal elements appear normally formed and without accumulation of trabecular or cortical bone.
Fig 2.
Low and high power histology of adjacent sections (A, B and C, D) of OCst-KO (A, B, E, F) and WT (C D, G, H) 6-week-old mouse distal femur.
Glycol methacrylate, 3 μm sections were stained histochemically for TRAP (B, D, E, G) and some were counterstained with toluidine blue (A, C, F. H). At low power, overall appearance was highly similar, with growth plates normal (purple, wavy band in A and C, asterisks in B and D), similar trabecular size and thickness in both the primary and secondary ossification centers, and open marrow space (m) in the diaphysis. TRAP stains (B, D) show overall similar distribution of osteoclasts, including along the chondroosseous junction, among the trabeculae of the primary ossification center, and notably, along the periosteal surface (arrows in B and D), where the bone is being removed to maintain the flared shape of the metaphysis. At higher power (E, F, G, H), only mononuclear osteoclasts are present in OCSt-KO mice. A row of such cells is indicated in A by arrows, attached to trabecular bone (B) in the primary spongiosa. In contrast, wild type mice had typical, multinucleated osteoclasts in this area, and individual nuclei (visible by lack of overlying TRAP label in C), are indicated by asterisks. Blinded observers consistently identified OCSt-KO and WT sections based solely on the presence or absence of multinucleated osteoclasts. Bar in A = 500 μm in A, B, C, D. Bar in E = 10 μm in E and G; 7.3 μm in F and H. B = bone, C = cartilage core.
Fig 3.
TRAP-positive osteoclast measurements on bone sections.
Images were obtained with a 5X objective of distal femora of WT and OCSt-KO mice at 6 weeks post-partum and TRAP-positive total area, number of TRAP-positive sites, and mean area of those sites were measured. A, A typical whole image of distal femur metaphysis, stained histochemically for TRAP (red) without counterstain. The region outlined in A is enlarged in B. C. shows the same area of the section after color matching was performed to select TRAP-positive sites in the section (now black). Finally, the images were made binary (D), leaving only TRAP-positive (black) and TRAP-negative (white) areas for analysis. Bar in A = 100 μM, bar in B = 50 μm in B, C, and D. E. The mean total TRAP-positive area per micrograph differed significantly between WT and OCSt-KO (left; *P< 0.05). The total count of TRAP-positive sites was not different between genotypes (middle); however, the mean area of each TRAP-positive sites was significantly lower in OCSt-KO (*P < 0.02). Results from 3 or 4 sections per animal, and 3 animals per genotype were pooled for analysis.
Fig 4.
Bone formation/resorption indicators were normal in OCSt-KO mice.
Serum levels of bone resorption markers collagen C-terminal peptides (CTX, left), and TRAP (middle), as well as the bone formation marker, osteocalcin (right), were measured in OCSt-KO and WTmice at 6 weeks post-partum. No significant differences in any of these parameters were found, consistent with normal rates of bone resorption and formation. Triplicate assays were run on samples from 9 (CTX), 5 (TRAP), and 9 (osteocalcin) WT mice, and on 4 (CTX), 3 (TRAP), and 3 (osteocalcin) OCSt-KO mice.
Fig 5.
Rescue of OCSt-KO BMMC by lentiviral transduction.
TRAP stain shows BMMC from WT mice transduced with GFP (A: WT), from OCSt-KO mice transduced with GFP (B: OC-STAMP KO), and from OCSt-KO mice transduced with OC-STAMP:GFP (C: Rescue) following 6 days of culturing in the presence of RANKL. Many large, multinucleated osteoclasts are seen in WT cells, whereas none are present in the KO cells. Fusion is rescued by transduction with OC-STAMP. Scale bar in B = 200 μm. Boxed areas in A, B, and C upper panels are shown at higher magnification below. Some individual nuclei within multinucleated cells are indicated by arrows in the WT and Rescue panels. D. BMMC were transduced and cultured as above on HA coated plates, and the plates were scanned after 6 days of culture. Resorbed HA appears as black. E. Quantitation of resorbed area shows a roughly 3.5-fold decrease of resorption activity in KO vs. WT, whereas the rescued cells had their activity mostly restored. Mean + s.d. is shown, n = 3, *P < 0.002; **P < 0.005. F. TRITC-labeled phalloidin shows large actin rings on dentine disks in WT (left) and rescued (right) cells. KO mononuclear osteoclasts (middle) also made actin rings, but they were much smaller, consistent with small cell size. Scale bar in F = 100 μm.
Fig 6.
Viability, secretion, and gene expression.
A. Viability. BMMC from wild type (WT) or OCSt-KO (KO) were virally transduced, cultured with or without RANKL for 6 days as indicated, and MTT assays for cell viability were performed. Transduction with either GFP or OC-STAMP-GFP (OCSt:GFP) lentivirus had no effect on viability. Mean + s.d., n = 3. B. TRAP secretion. Supernatants from cells cultured as in A were analyzed for secreted TRAP enzyme. No significant difference in TRAP secretion was seen between wild type (WT) and OCSt-KO (KO) cells, whether transduced with GFP or with OC-STAMP fused to GFP (OCSt:GFP). Mean + s.d., n = 3. C-H. mRNA expression of osteoclast markers. Primary BMMC cells were virally transduced and were cultured for 0, 3, and 6 days in the presence of RANKL, and Q-PCR analyses for osteoclast mRNAs were performed, as indicated. Note that no OC-STAMP mRNA was detected in OCSt-KO (KO) cells at any time point unless the viral vector encoded OC-STAMP (OCSt:GFP), whereas no other osteoclast markers were significantly affected; n = 3.
Fig 7.
Peripheral localization of OC-STAMP and topology of N- and C-termini.
A-C. HEK293T cells were transfected with either GFP or OC-STAMP-GFP vectors (OCSt:GFP). A. Western blot with anti-GFP antibody shows expected molecular weights (marker positions indicated at left). B and C, GFP fluorescence. In B, endogenous GFP fluorescence is seen throughout the cell, as expected for a cytoplasmic protein. In C, OC-STAMP-GFP signal is mainly confined to the cell periphery, as expected for a plasma membrane-localized transmembrane protein. Scale bar = 50μm. D-F. HEK 293 cells were transfected with tagged constructs indicated at top, permeabilized as indicated at left (none, plasma membrane selectively with digitonin, or all membranes with Triton X-100), and probed with anti-tag antibodies. With membranes intact, anti-FLAG antibody could not access the N-terminus, nor could anti-GFP access the C-terminus, whereas both antibodies could access their antigens when the plasma membrane or all membranes were permeabilized. For the GFP tag, separate green and red channels are also shown in small panels. GFP endogenous signal (green) vs. that for anti-GFP (red) varied in intensity in some cells, but the same cells were positive for both. Scale bars = 10 μm.
Fig 8.
A. HEK293 cells were transfected with V5-tagged wild type (WT) OC-STAMP or glycosylation-deficient (N162D) OC-STAMP. Wild type-transfected cell extracts were either untreated or digested with N-glycanase (N-Glyc’). Extracts were blotted and probed with anti-V5 antibody. Arrowheads indicate 2 bands, an upper, glycosylated form approximately 3kDa higher than the lower band at 50 kDa. Both N-glycanase-treated and the N >D mutation show only the low molecular weight form. Some of the overexpressed WT protein escapes glycosylation, giving 2 bands. B. BMMC from OCSt-KO mice were transduced with lentiviral vectors encoding wild type (WT) or (N162D) OC-STAMP fused to GFP. Cells were cultured for 6 days in RANKL, and extracts were blotted and probed with anti-GFP (upper panel) or anti-α-tubulin (lower panel). The glycosylated WT OC-STAMP appears to have much greater stability. C-F. TRAP enzyme cytochemistry. Wild type (WT) or OCSt-KO BMMCs were differentiated with RANKL for 6 days in 96-well plates. They were transduced with either GFP (C, D), OC-STAMP-GFP fusion protein (OCSt:GFP; E), or OC-STAMP-GFP fusion carrying the N162 mutation D (OC-StN>D; F). Fusion was rescued in the knockout cells by either the glycosylated or the non-glycosylated form of OC-STAMP. Bar in D = 200 μm.
Table 1.
Conservation of glycosylation motif in terrestrial vertebrates.
Fig 9.
Combining experimental results and topology predictions, the diagram shows current understanding of OC-STAMP topology with respect to the plasma membrane (PM). N- and C-termini are indicated, along with the position of the N-terminal FLAG and C-terminal GFP tags, and glycosylated residue N162. TM = transmembrane helix, IL = intracellular loop, EL = extracellular loop. A consensus mammalian core oligosaccharide is shown attached to N162: blue = GlcNac, green = mannose, yellow = galactose, red = Neu5Gc.
Table 2.
Topology of mouse OC-STAMP.