Severe growth retardation and early lethality in mice lacking the nuclear localization sequence and C-terminus of PTH-related protein

Abstract

Parathyroid hormone (PTH) plays a central role in the regulation of serum calcium and phosphorus homeostasis, while parathyroid hormone-related protein (PTHrP) has important developmental roles. Both peptides signal through the same G protein-coupled receptor, the PTH/PTHrP or PTH type 1 receptor (PTH1R). PTHrP, normally a secreted protein, also contains a nuclear localization signal (NLS) that in vitro imparts functionality to the protein at the level of the nucleus. We investigated this functionality in vivo by introducing a premature termination codon in Pthrp in ES cells and generating mice that express PTHrP (1–84), a truncated form of the protein that is missing the NLS and the C-terminal region of the protein but can still signal through its cell surface receptor. Mice homozygous for the knock-in mutation (Pthrp KI) displayed retarded growth, early senescence, and malnutrition leading postnatally to their rapid demise. Decreased cellular proliferative capacity and increased apoptosis in multiple tissues including bone and bone marrow cells were associated with altered expression and subcellular distribution of the senescence-associated tumor suppressor proteins p16^INK4a and p21 and the oncogenes Cyclin D, pRb, and Bmi-1. These findings provide in vivo experimental proof that substantiates the biologic relevance of the NLS and C-terminal portion of PTHrP, a polypeptide ligand that signals mainly via a cell surface G protein-coupled receptor.

Intranuclear transport of numerous polypeptide ligands has been reported, yet the function of these ligands at the level of the nucleus and its biological relevance in the in vivo setting remain in question (1–3). The parathyroid hormone-related protein (Pthrp) gene encodes a nuclear localization signal (NLS) within the 87 to 107 region of the mature protein product and contains at least two translational initiation sites, one that generates a conventional signal peptide and one that disrupts it (4, 5). These features allow PTHrP either to be secreted in a paracrine/autocrine fashion or to be retained within the cytosol and to be translocated to the nucleus (4, 6, 7). Alternative potential mechanisms of PTHrP entry into the cytosol have also been described (8–10).

The similarity of the N terminus of PTHrP to that of parathyroid hormone (PTH), the major hormone regulator of calcium and phosphorus homeostasis, enables PTHrP to share the signaling properties of PTH by interacting with the common PTH/PTHrP or PTH type 1 receptor (PTH1R), a member of the G protein coupled receptor family B. Previous studies have clearly documented the indispensable biological importance of PTHrP/PTH1R signaling in skeletal (11–15) and mammary gland (16, 17) development. Additional endocrine and paracrine functions have been ascribed to the mid-region and C-terminal region (107 to 139) of the molecule (18, 19). In vitro studies have indicated that PTHrP displays other functions largely relating to an intracrine signaling role in the nucleus/nucleolus (4, 20, 21) and that its subcellular distribution is cell-cycle dependent (22) in that PTHrP is targeted to the nucleolus of cells in G₁.

In the present study, we examined the potential biological relevance of the nuclear localization of PTHrP in vivo by inserting a premature termination codon (TGA) in the murine Pthrp, and generating a “knock-in” (KI) mouse expressing PTHrP (1–84), a form of the protein that lacks the NLS and C-terminal region. Using this approach, Pthrp is expected to be expressed in relevant cells and at physiologic levels. In addition, the truncated PTHrP protein should be processed and secreted freely within tissues to bring about the appropriate paracrine/autocrine effects by interacting with PTH1R, which localizes at the plasma membrane but would be devoid of any potential nuclear actions.

Results

Generation of a “Knock-In” (KI) Mouse Expressing PTHrP (1–84).

We first targeted the Pthrp locus in ES cells by introducing a premature termination codon at amino acid position 85 of the encoded mature protein [see supporting information (SI) Fig. S1 A–D]. Appropriately targeted ES cells from two clones containing the mutated Pthrp gene were microinjected into 3.5-day C57BL/6 blastocysts and then transferred into uteri of 2.5-day post coitus pseudopregnant CD1 mice to generate chimeric animals. Extensively chimeric male mice were crossed to C57BL/6 female mice and heterozygous offspring were identified by PCR of tail genomic DNA (Fig. S1E). They, in turn, were bred to obtain animals homozygous for the mutated Pthrp allele (Pthrp KI mice). Absence of PTHrP (1–84) truncated mutant from the nucleus was confirmed in mouse embryonic fibroblasts (MEFs) derived from these mice and appropriate expression levels were verified (Fig. S2 A-D).

Pthrp KI Mice Exhibit Growth Retardation and Premature Aging.

The phenotype of Pthrp KI mice was distinct from that of homozygous Pthrp-null mice (11, 12, 23). At birth, Pthrp KI mice were similar in weight to wild-type littermates. However, by three days postpartum they failed to grow relative to their wild-type littermates (Fig. 1A) and died by two to three weeks of age. Serum levels for calcium, phosphorus, and PTH were normal at two weeks of age (Fig. S2 E-G). The Pthrp KI mice were slightly smaller than wild-type littermates without markedly shorter limbs at birth, a distinguishing feature of the homozygous null mice (Fig. 1B). At 2 weeks, the Pthrp KI mice exhibited marked phenotypic changes indicative of premature aging including an unstable gait, cachexia, osteopenia with kyphosis (Fig. 1C), and a profound decrease in fat deposition (Fig. 1D). The brains of these animals were both smaller and edematous (Fig. 1E). The skin was thin with hyperkeratosis of the epidermis (Fig. 1F) and large blood vessels were often atrophic and therefore potentially more subject to rupture (Fig. 1G). Nevertheless, it was difficult to identify a single cause of death in our animals and indeed premature senescence often leads to death from multiorgan failure (24). The senescence biomarker β-galactosidase (25) was present in tissues such as kidney (Fig. 1H) and lung (data not shown), suggesting that these tissues undergo early onset senescence. This early senescence phenotype contrasts strikingly to the phenotypes of the Pthrp-null (11, 12) and those of the Pth1r-null mice (13).

Fig. 1.

Pthrp KI mice exhibit growth retardation and premature aging. (A) Growth curves of wild-type (WT) and Pthrp KI mice. (B) Whole mount skeletons of newborn WT and Pthrp KI mice stained with alcian blue (for cartilage) and alizarin red (for calcified tissue). (C) Surviving KI mice at 14 days of age show osteopenia and severe kyphosis (arrow). (D) KI mice at day 14 show loss of body mass, muscle atrophy, and a profound decrease in adipose tissue (Ad). (E) Representative micrographs of brains showing diffuse swelling of the tissue with complete obliteration of the sulci in the KI mice. (F) Representative micrographs of skin from Pthrp KI mice showing thinner skin (green line) with hyperkeratosis of the epidermis (blue line) (magnification, ×100). (G) Representative micrographs of common carotid arteries (magnification, ×400). (H) Representative micrographs of renal tissue sections showing endogenous β-galactosidase activity in renal cortical tubules from Pthrp KI mice, but not from WT mice (magnification, ×400). (I) Contact radiographs (Left) and microCT 3D reconstruction (Right) of the mandibles from WT and Pthrp KI mice (Left and Right specimen in each panel, respectively). (J) Western blot analysis of skeletal muscle extracts for LC3-I and II. ß-Tubulin was used as loading control.

Influence of Nutrition on the Phenotype of the Pthrp KI Mice.

We next assessed the potential influence of nutrition on the phenotype of the Pthrp KI mice. Stomach milk weights (Fig. S3A) were similar during the first two days but lower with time, although milk intake relative to body weight was not significantly reduced after day 1 (Fig. S3B). Blood glucose levels were reduced (Fig. S3C) in keeping with previous observations in other early senescent models (26–28). Interestingly, in contrast to the Pthrp-null mice (29), tooth eruption occurred in the Pthrp KI mice (Fig. 1I).

We then sought to compare the presence of autophagy in 4-day-old wild-type fed mice, 4-day-old wild-type mice that had been starved for 24 h, and 4-day-old Pthrp KI fed mice. Immunoblots of microtuble-associated protein light chain 3 (LC3-I), an autophagosome marker protein (30, 31), were performed in skeletal muscle, a major site of the manifestation of malnutrition. LC3-I was only modestly converted to membrane bound (lipidated) LC3 (LC3-II) in fed wild-type mice but was markedly converted in the starved wild-type mice (Fig. 1J). In contrast, levels of LC3-II relative to LC3-I in the Pthrp KI mice were greater than in the fed wild-type, but substantially lower than in the starved mice (Fig. S3 D and E). Although undernourishment may have contributed in part to the phenotype, these levels in Pthrp KI mice were comparable to the degrees of autophagy observed in other progeroid mouse models (32).

Skeletal Growth Retardation Caused by Impairment in Endochondral Bone Formation.

In E18.5 Pthrp KI mice, growth plates were reduced in size with narrower proliferative zones (Fig. 2 A and B), with no evidence of the markedly disrupted architecture in the hypertrophic zone characteristic of the growth plates of Pthrp-null mice (12). In 2-week-old Pthrp KI mice, long bones were markedly shorter (Fig. 2 C and D) and osteoporotic, as determined by decreased bone mineral density (Fig. 2E), while epiphyseal volumes were reduced, resulting in small growth plates, as verified by microCT imaging and histology (Fig. 2 F-H). Cartilaginous matrix mineralization, however, was appropriate (Fig. 2 I and J). The proliferation of chondrocytes, as determined by immunostaining for proliferating cell nuclear antigen (PCNA), was profoundly diminished in the KI mice (Fig. 2 K and L). Staining also decreased for type X collagen, a marker of terminal chondrocyte differentiation (Fig. 2 M and N). However, the pattern of cellular differentiation within the growth plate was normal. This therefore reflected a reduction in differentiated chondrocytes, which was likely secondary to their decreased proliferative capacity. This phenotype, although resulting in skeletal growth retardation, stood in striking contrast to the chondrodysplastic Pthrp-null long bones, where the disorganized hypertrophic zone is secondary to premature and inappropriate differentiation of chondrocytes (12).

Fig. 2.

Skeletal growth retardation caused by impaired growth plate chondrocyte proliferation. (A) Representative micrographs of H&E stained sections of the proximal ends of tibiae from E18.5 wild-type (WT) and Pthrp KI mice (magnification, ×100). Blue lines represent the respective proliferating zones. (B) The length measurements of the proliferating zone. (C) Representative contact radiographs of the femurs of WT and Pthrp KI mice at 2 weeks of age. (D) Femoral length measurements. (E) BMD measurements. (F) Representative frontal views of microCT 3D reconstruction of the proximal end of tibiae. (G) Epiphyseal volume of the proximal ends of tibiae. (H) Width of the cartilaginous growth plate. (I) Undecalcified sections of tibiae stained by the von Kossa procedure (magnification, ×200). (J) Mineralized area, percentage of growth plate. (K) Paraffin-embedded sections of tibiae from WT and KI mice stained immunohistochemically for PCNA (arrowheads) (magnification, ×400). (L) Number of PCNA-positive chondrocytes as a percentage of total chondrocytes, as determined by image analysis. (M) Immunostaining for type X collagen (arrowheads) (magnification, ×100). (N) Width of type X collagen-positive hypertrophic zone of growth plates. Data shown represent mean ± SE of five animals per group. **, P < 0.01; ***, P < 0.001 in the KI mice relative to the wild-type mice.

Premature Osteoporosis Results from Defective Osteoblastic Bone Formation.

The trabecular bone of E18.5 Pthrp KI mice showed reduced osteoblasts (Fig. 3A-C). In 2 week old mice, the longitudinal and cross sectional views from 3-D microCT reconstructions of long bones further confirmed the osteoporotic phenotype of the Pthrp KI mice (Fig. 3D). Trabecular and cortical bone volumes, and trabecular number and thickness were decreased, whereas trabecular spacing was increased (Fig. 3 E-I). Tartrate resistant acid phosphatase (TRAP) activity and the number of osteoclasts were decreased in mutant bones (Fig. 3 J and K). The immunopositive area for osteoblastic PTH1R was also dramatically decreased in long bone sections from Pthrp KI mice (Fig. 3 L and M) and associated with a marked reduction in osteoblast number (Fig. 3 N and O) and protein expression of Cbfa1 and Pth1r (Fig. 3P). We therefore concluded that decreased osteoblastic bone formation was the major cause of the osteoporotic phenotype in the Pthrp KI mice. This bone phenotype was not secondary to altered external humoral or nutritional factors, as determined by bone transplantation experiments (Fig. S4 A and B).

Fig. 3.

Severe premature osteoporosis resulted from defects of osteoblastic bone formation. (A) Representative micrographs of hematoxylin and eosin (H&E) stained sections of the metaphyseal region of tibiae from E18.5 wild-type (WT) and Pthrp KI mice (magnification, ×400). (B) Osteoblast number/tissue area (N.Ob/T.Ar), (C) Osteoblast surface/bone surface ratio (Ob.S/BS). (D) Representative longitudinal (top) and cross sections (bottom) of 3D reconstruction of proximal end of tibiae at 2 weeks of age. Quantitative histomorphometry for (E) bone volume/total volume (BV/TV), (F) cortical bone volume (Ct.V.), (G) trabecular number (Tb.N), (H) trabecular thickness (Tb.Th), and (I) trabecular separation (Tb.Sp). (J) Representative micrographs of tibial sections from the WT and KI mice stained histochemically for TRAP (magnification, ×200), (K) Osteoclast number/tissue area (N.Oc/T.Ar), (L) Representative micrographs of tibial sections from the WT and KI mice stained immunohistochemically for Pth1r (magnification, ×400), (M) Pth1r positive area as percent of tissue. (N) Representative micrographs of tibial sections from the WT and KI mice stained with H&E (magnification, ×400). (O) Osteoblast number/tissue area (N.Ob/T.Ar). (P) Western blot of long bone extracts for the expression of Cbfa1 and Pth1r. ß-tubulin was used as loading control for Western blots. Data shown represent mean ± SE of five animals per group. ***, P < 0.001 in the KI mice relative to wild-type littermates.

Reduced Cell Proliferation and Increased Cellular Apoptosis in Pthrp KI Tissues.

In addition to the reduced numbers of proliferative chondrocytes and osteoblasts in situ, in utero, we observed a profound decrease in PCNA positive stem/progenitor cells in the subventricular zone and the hippocampus in brains from E18.5 Pthrp KI mice (Fig. 4 A and B). In 14-day-old mutant mice, bone marrow cells showed significantly lower incorporation of BrdU compared to their wild-type littermates (40.0% vs. 66.3%) (Fig. 4 C and D). Moreover, TUNEL-positive cells and annexin V⁺/propidium iodide-negative (PI⁻) cells were increased in the thymus and spleen of Pthrp KI mice compared to wild-type littermates (Fig. 4 E-J). Consistent with increased cellular apoptosis in Pthrp KI mice, expression of the proapoptotic protein Bax was dramatically increased, whereas that of the anti-apoptotic protein Bcl-2 was decreased (Fig. 4K), suggesting that the observed growth retardation and in vivo senescence phenotypes were associated with inhibition of cell proliferation and stimulation of cellular apoptosis.

Fig. 4.

The Pthrp KI phenotype is associated with inhibition of cell proliferation and stimulation of cell apoptosis. (A) Representative micrographs of sections from the subventricular zone (SVZ) and the hippocampus (HP) of brains from E18.5 WT and KI mice immunostained for PCNA (brown color indicated by arrowheads; magnification, ×400). (B) PCNA positive cell numbers in SVZ and HP (number/per field). (C) Bone marrow cells from 14-day-old Pthrp KI mice show decreased incorporation of BrdU (red) compared to those from WT mice. (D) Quantitative assessment of BrdU incorporation using flow cytometry (blue profile for negative control, red profile for BrdU positive cells). (E) Representative micrographs of thymus sections from WT and KI mice stained for apoptotic cells using the TUNEL technique (red, magnification, ×1000). (F) Flow cytometry analysis of apoptotic thymocytes (Annexin-V positive (+)/PI negative (−) cells; blue profile for PI positive cells, red profile for Annexin-V positive cells). (G) Representative spleen sections from WT and KI mice stained for apoptotic cells using the TUNEL technique (red, magnification, ×1000). (H) Flow cytometry analysis of apoptotic spenocytes (blue profile for PI positive cells, red profile for Annexin-V positive cells). (I) The percentage of apoptotic cells in thymus and spleen as determined by TUNEL assay were quantified by image analysis and are presented as mean ± SE of triplicate determinations. (J) The percentage of Annexin-V positive (+)/PI negative (−) cells in thymus and spleen was quantified from flow cytometry analysis. Data shown represent mean ± SE from five animals per group. *, P < 0.05; **, P < 0.01; ***, P < 0.001 for the Pthrp KI samples relative to those from wild-type littermates. (K) Western blot analysis of thymus and spleen extracts for Bax and Bcl-2. ß-Tubulin was used as loading control.

Altered Senescence-Associated Tumor Suppressor Genes and Oncogene Expression in Pthrp KI Mice.

In view of the fact that senescence is the final phenotypic state of decreased cell proliferation often mediated by increased expression and activation of tumor suppressor genes, we examined the expression of p16^INK4a and p21 as well as that of the oncogene Bmi-1, a member of the Polycomb/trixthorax group (Pc-G/trx-G) proteins (33, 34). Indeed, expression of p16^INK4a and p21 was significantly increased, while that of Bmi-1 was reduced in tissues (including bone, thymus, and spleen) and in MEFs derived from Pthrp KI mice compared to wild-type counterparts even when mutant MEFs were cultured under conditions of full nutritional supplementation (Fig. 5A). Furthermore, nuclear localization of p16^INK4a and p21 was increased in MEFs from KI mice (Fig. 5B). In contrast, Bmi-1 was detected in the nuclei of MEFs from wild-type but not in nuclei of MEFs derived from Pthrp KI mice. In addition, levels of Cyclin D, Cdk4, and Cdk6 and the phosphorylated form of the retinoblastoma gene product (pRb), all of which are essential for cell proliferation and have been found to be altered in senescent states (35) were all reduced (Fig. 5C).

Fig. 5.

Altered expression and subcellular distribution of senescence-associated tumor suppressor proteins and oncogenes. (A) Western blot analysis of long bone and MEF extracts for p16^INK4a, p21, and Bmi-1. ß-Tubulin was used as loading control. (B) MEFs cultured for 14 days in DMEM with 10% FCS were stained immunocytochemically for p16^INK4a, p21, and Bmi-1 (red; magnification, ×1000). (C) Western blot analysis of long bone extracts for Cyclin D, Cdk4, Cdk6, and phosphorylated Rb (pRb). ß-Tubulin was used as loading control. (E) Proposed regulation of proliferation and senescence in cells that expresses either PTHrP (1–139) (Left) or PTHrP (1–84) which lacks the NTS and the C terminus (CT) (Right). In the absence of NTS and C terminus PTHrP, Bmi-1 remains inactive in the cytoplasm, leading to increased p16^INK4a (as well as p21) levels, G₁ cell-cycle arrest and senescence.

Discussion

Our findings indicate that in vivo deletion of the NLS and C-terminal region of PTHrP leads to growth retardation and early senescence by altering expression patterns and subcellular distribution of proliferative and senescence-related genes in multiple tissues. Although malnutrition manifested by autophagy may have contributed in part to the observed phenotype, the Pthrp KI mouse is a progeroid mouse model as deposition of β-galactosidase, regarded as a senescence marker even in the presence of autophagy (36), was detected in its tissues.

Interactions of the PTHrP Domains.

Previous in vivo studies have reported that Pthrp null mice (11, 12) as well as Pth1r null mice (13), demonstrate dysplastic long bone formation and early lethality (i.e., in utero or within one to two days following birth). Targeted overexpression of PTHrP in chondrocytes of Pthrp null mice (37) largely eliminated the growth plate pathology as did targeted overexpression of PTH1R in chondrocytes of Pthrp (38) and Pth1r null mice (39). Both models displayed reduced survival despite improvement in long bones, however, indicating that the defect in viability does not result from the skeletal abnormalities, but detailed descriptions of extra-skeletal pathology have not been available. The presence of a markedly distinct phenotype in the long bones and teeth of our Pthrp KI mice supports the unique action of the nuclear localizing domain of PTHrP in skeletal pathology. The early lethality in Pth1r null mice (13) as well as in our mice lacking the nuclear localizing/C-terminal domains, may indicate that unique mechanisms may exist for PTH1R and for the nuclear localizing domain to promote optimal survival, and the absence of both regions, in the Pthrp null mice may co-operatively contribute to the more rapid death of these animals than of the mice lacking the nuclear localizing/C-terminal domains. In this respect, the latter PTHrP domains have been implicated in enhancement of proliferation (6, 20, 21), a function reported to require both an intact NLS and an intact C-terminal region (20), and in inhibition of apoptosis (4, 6, 40, 41) in several cell systems in vitro. In contrast, the amino (N) terminal domain has been reported to either stimulate (42) or inhibit (6, 43) proliferation depending on the in vitro system used, and to also stimulate (44) or inhibit (45) apoptosis. The biological actions of PTH1R apparently involved in longevity (as well as in growth plate and tooth eruption) most likely reside in local actions of the N terminus of PTHrP rather than in circulating PTH, in view of the fact that Pth null mice do not exhibit a reduced life span (23) and PTH levels were normal in our model. These distinct actions of the N- terminal domains of PTH and PTHrP may reflect conformational selectivity for the PTH1R, as recently described (46).

Modulation of Cell Cycle and Apoptosis by PTHrP.

In vitro studies investigating the cellular mechanisms responsible for the mitogenic checkpoint release by nuclear PTHrP have shown that in vascular smooth muscle cells the NLS, together with the C-terminal region of PTHrP, is translocated to the nucleus of cells in G₁ (22). The nuclear presence of PTHrP appears to trigger pRb phosphorylation and release of G₁/S arrest and thereby cell cycle progression (21). In this setting, cyclin E/Cdk-2 kinase activity is markedly increased by PTHrP as a result of marked PTHrP-induced proteasomal degradation of p27^kip1 (47). In contrast, we show here that in Pthrp KI cells, one consequence of the absence of nuclear PTHrP action in vivo is increased p21 levels, which in turn would lead to inhibition of cyclin E/Cdk2 and cyclin D1/Cdk4/Cdk6 activities (shown to be reduced in our studies) and to cell-cycle arrest in G₁ phase.

PTHrP has also been shown to bind to RNA (48). In view of evidence that some rDNA transcription occurs largely at the border between the fibrillar center and the dense fibrillar component of the nucleolus (49) where PTHrP localizes in vivo (4), it is conceivable that it partakes in regulating a variety of nucleolar functions required to support cell growth and division, and coordination of cellular stress responses (50).

Bmi-1 is required for the maintenance of adult stem cells because it promotes cell proliferation and suppresses genes that induce cellular senescence (33) and cell death (34). Its action is due in part to its ability to suppress expression of proteins that inhibit cell cycle progression such as p16^INK4a (33). Expression of p16^INK4a rises markedly with aging in many tissues (51), inhibits cyclin D-dependent kinases, and prevents phosphorylation of Rb, thus limiting proliferation and self-renewal (50). This reduces the reservoirs of self-renewing tissue stem cells required to regenerate lost or damaged cells with aging. We have shown here that p16^INK4a is highly expressed in tissues from Pthrp KI mice and that phosphorylated Rb is reduced suggesting that Bmi-1 function is impaired.

Nuclear translocation of Bmi-1 is necessary for its function (52, 53), and Bmi-1 that retains its cytoplasmic localization is inactive (52). In our studies, Bmi-1 localized to the nucleus in wild-type cells, but not in cells derived from Pthrp KI mice, most likely resulting in failure to suppress p16^INK4a expression in Pthrp KI MEFs and tissues. It is probable that the nucleocytoplasmic shuttling ability of PTHrP (54, 55), which is determined by its nuclear localizing domain, is involved in the nuclear shuttling of proteins such as Bmi-1 and disruption of this mechanism results in inactive Bmi-1 (Fig. 5D). Further studies will be required to delineate the importance of these actions in the regulation of cellular proliferation and senescence.

Biological Implications.

We propose that the Pthrp KI mouse is the first laboratory animal model with a distinct phenotype arising from the inability of a peptide hormone to act within the nucleus while maintaining its role as a secreted ligand. Undoubtedly, these studies do not shed light on whether the observed phenotypic alterations are exclusively a consequence of loss of C-terminal PTHrP action at the level of the nucleus or elsewhere in the cell. Our conclusions, however, are corroborated by in vitro observations showing that activation of cell proliferation requires both an intact NLS and an intact C terminus and that deletion of the NLS prevents nuclear entry and slows proliferation (20). The demonstration here of the functionality of the NLS and C terminus of PTHrP at the tissue and organismal level now adds further credence to the biologic relevance of nuclear transport and function of polypeptide ligands. Identifying bona fide targets within the nucleus should now provide important clues to the relationship between this process and aging.

Materials and Methods

Genotyping of Mice.

Genomic DNA was isolated from tail fragments by standard phenol/chloroform extraction and isopropyl alcohol precipitation. PCR was conducted to determine the genotype at the Pthrp locus. DNA was amplified with forward primer 5′-GCTGTGTCTGAACATCAGCTAC-3′ and reverse primer 5′-ATGCGTCCTTA-AGCTGGGCTC-3′. Cycling conditions were 94 °C for 30 sec, 60 °C for 30 sec, and 72 °C for 45 sec (35 cycles), followed by an extension at 72 °C for 10 min. PCR products were digested with BstEII at 60 °C for 2 h and analyzed by agarose gel electrophoresis. The DNA from wild-type mice gave only one band of 424-bp while that from homozygous KI mice produced 2 bands of 258-bp and 166-bp. The heterozygous mice produced 3 bands of 424-, 258-, and 166-bp.

Statistical Analysis.

Data from image analysis are presented as means ± SE. Statistical comparisons were made using a two-way ANOVA, with P < 0.05 being considered significant.