Klotho Protein Deficiency Leads to Overactivation of μ-Calpain*

The klotho mouse is an animal model that prematurely shows phenotypes resembling human aging. Here we report that in homozygotes for the klotho mutation (kl −/−), αII-spectrin is highly cleaved, even before the occurrence of aging symptoms such as calcification and arteriosclerosis. Because αII-spectrin is susceptible to proteolysis by calpain, we examined the activation of calpain in kl −/− mice. m-Calpain was not activated, but μ-calpain was activated at an abnormally high level, and an endogenous inhibitor of calpain, calpastatin, was significantly decreased. Proteolysis of αII-spectrin increased with decreasing level of Klotho protein. Similar phenomena were observed in normal aged mice. Our results indicate that the abnormal activation of calpain due to the decrease of Klotho protein leads to degradation of cytoskeletal elements such as αII-spectrin. Such deterioration may trigger renal abnormalities inkl −/− mice and aged mice, but Klotho protein may suppress these processes.

The klotho mouse is an animal model that prematurely shows phenotypes resembling human aging. Here we report that in homozygotes for the klotho mutation (kl ؊/؊ ), ␣ II -spectrin is highly cleaved, even before the occurrence of aging symptoms such as calcification and arteriosclerosis. Because ␣ II -spectrin is susceptible to proteolysis by calpain, we examined the activation of calpain in kl ؊/؊ mice. m-Calpain was not activated, but -calpain was activated at an abnormally high level, and an endogenous inhibitor of calpain, calpastatin, was significantly decreased. Proteolysis of ␣ II -spectrin increased with decreasing level of Klotho protein. Similar phenomena were observed in normal aged mice. Our results indicate that the abnormal activation of calpain due to the decrease of Klotho protein leads to degradation of cytoskeletal elements such as ␣ II -spectrin. Such deterioration may trigger renal abnormalities in kl ؊/؊ mice and aged mice, but Klotho protein may suppress these processes.
The klotho (kl Ϫ/Ϫ ) mouse shows multiple phenotypes resembling human aging caused by the mutation of a single gene (1). This mutation is caused by the insertion of ectopic DNA into the regulatory region of the klotho gene. The klotho gene encodes a type I membrane protein that is expressed predominantly in the kidney and brain. The extracellular domain of Klotho protein consists of two internal repeats that share sequence similarity to the ␤-glucosidases of both bacteria and plants (1,2). As a result of a defect in klotho gene expression, the kl Ϫ/Ϫ mouse exhibits multiple age-associated disorders, such as arteriosclerosis, osteoporosis, skin atrophy, pulmonary emphysema, short life span, and infertility. However, the mechanism by which the klotho gene product suppresses the aging phenomena has not been identified. Analysis of the pathophysiology of kl Ϫ/Ϫ mice is expected to give clues not only to understanding the mechanisms of individual diseases associated with aging but also the relationship between these mechanisms during human aging.
Non-erythroid spectrin is a heterodimeric actin-binding pro-tein that consists of ␣ II -and ␤ II -spectrin and is usually found on the cytoplasmic side of the plasma membrane (3,4). It is thought to participate in the establishment and maintenance of cell polarity, shape, and receptor distribution (5). Recently, it was proposed that spectrin retained and stabilized various proteins at specific regions on the cell surface (6 -9). ␣ II -Spectrin has been shown to be cleaved by calpain and/or caspase during apoptosis and necrosis (10 -13). Calpain, a calcium-dependent cytosolic cysteine protease, is involved in many physiological and pathological processes (14 -16). Calpain mediates proteolysis of various cellular proteins, including cytoskeletal proteins, and causes irreversible cell damage (10 -13, 17, 18). Thus, calpain overactivation may contribute to the pathology of cerebral and cardiac ischemia, Alzheimer's disease, arthritis, and cataract formation (19,20). Calpain has been shown to be regulated by both calcium ion and calpastatin (16). Two types of isozymic calpain, -calpain and m-calpain, are ubiquitously distributed in mammalian cells. The former is activated by micromolar concentrations of calcium and the latter is activated by millimolar concentrations of calcium. Calpastatin is an endogenous inhibitor specific for calpain, but is slowly degraded by calpain (21). Here, we report the cleavage of ␣ II -spectrin due to the continuous activation of -calpain in kl Ϫ/Ϫ mice. Furthermore, we also observe similar phenomena in normal aged mice.

EXPERIMENTAL PROCEDURES
Preparation of Mouse Tissue Extracts-Kidneys were obtained from 2-and 3-week-old kl ϩ/ϩ , kl ϩ/Ϫ , and kl Ϫ/Ϫ mice and from 4-week-old and 29-month-old C57BL/6 mice. Brain, lung, heart, liver, and kidney were obtained from 4-week-old and 8-week-old kl ϩ/ϩ and kl Ϫ/Ϫ mice. Tissue samples were homogenized with 9 volumes (weight/volume) of 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 250 mM sucrose. After centrifugation at 900 ϫ g for 10 min, the supernatant was subjected to ultra centrifugation at 100,000 ϫ g for 1 h. The supernatants and precipitates were used as the cytosolic fraction and microsomal membrane fraction, respectively. Protein concentration was determined by BCA assay (Pierce). All experimental procedures using laboratory animals were approved by the Animal Care and Use Committee of Tokyo Metropolitan Institute of Gerontology. All efforts were made to minimize the number of animals used and their suffering.
Amino Acid Sequencing of 280-kDa Protein-Kidney microsomal fraction (250 g) from 4-week-old kl ϩ/ϩ and kl Ϫ/Ϫ mice was subjected to SDS-PAGE under reducing conditions followed by staining with Coomassie Brilliant Blue R-250. A protein band of ϳ280 kDa was excised and treated with 0.1 g of Achromobacter protease I (lysylendopeptidase) at 37°C for 12 h in 0.1 M Tris-HCl, pH 9.0, containing 0.1% SDS and 1 mM EDTA (22). The peptides were separated on columns of DEAE-5PW (1 ϫ 20 mm; Tosoh, Tokyo, Japan) and CAPCELL PAK C18 UG120 (1 ϫ 100 mm; Shiseido, Tokyo, Japan). Solvent A was 0.085% (v/v) trifluoroacetic acid in distilled water, and solvent B was 0.075% (v/v) trifluoroacetic acid in 80% (v/v) acetonitrile. The peptides were eluted at a flow rate of 30 l/min using a linear gradient of 1-60% solvent B. Selected peptides were subjected to Edman degradation using a Procise 494 cLC protein sequencer (Applied Biosystems, Foster City, CA) and to matrix-assisted laser desorption ionization time-offlight mass spectrometry on a Reflex MALDI-TOF (Bruker Daltonics, Billerica, MA) in linear mode using 2-mercaptobenzothiazole as a matrix.
Antibodies-Rabbit antibodies specific to the pre-and post-autolytic forms of -calpain (anti-pre-and anti-post-, respectively) were raised against synthetic peptides as described previously (23). Antibodies specific to the pre-and post-autolytic forms of m-calpain (anti-pre-m and anti-post-m, respectively) were produced using synthetic peptides corresponding to the N-terminal 21 residues (AGIAAKLAKDREAAEGLG-SHE) of the intact form and the N-terminal 6 residues (KDREAA) of the autolytic form, respectively. A cysteine residue was added to the C terminus of each peptide so that the antigenic peptide could be conjugated to keyhole limpet hemocyanin. The entire amino acid sequence and the autolytic cleavage site of human m-calpain were obtained from previous reports (24,25). Antibodies specific to the calpain-generated N-and C-terminal fragments of ␣ II -spectrin (136 and 148 kDa, respectively) were produced by the peptide antigens QQQEVY (anti-BDP-136) and GAMPRD (anti-BDP-148), respectively (see Fig. 2). A cysteine residue was added to the N terminus of QQQEVY peptide or was added to the C terminus of GAMPRD peptide. The amino acid sequence and the cleavage site in mouse ␣ II -spectrin by calpain were as determined by others previously (22,26). An antibody against domain IV of human calpastatin was produced using a synthetic peptide corresponding to residues 601-630 (AEHRDKLGERDDTIPPEYRHLLDDNGQDKP) (27) with a cysteine residue added to the C terminus. Rabbits were immunized with the antigenic peptide-keyhole limpet hemocyanin conjugates. Affinity purification of polyclonal antibodies was carried out using antigenic peptides immobilized on epoxy-activated Sepharose 6B (Amersham Biosciences, Buckinghamshire, UK). Anti-human-␣ II -spectrin polyclonal antibody C-20 from goat was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Klotho monoclonal antibody (KM2076) from rat was a generous gift from Kyowa Hakko Kogyo Co., Ltd (28).

RESULTS
Decrease of ␣ II -Spectrin in the Kidney of kl Ϫ/Ϫ Mice-To determine whether mouse homozygotes of the klotho gene mutation (kl Ϫ/Ϫ ) have a different pattern of proteins in the kidney, we examined the kidney microsomal fractions from 4-week-old mice by SDS-PAGE. A band of about 280 kDa was found to be significantly weaker in kl Ϫ/Ϫ mice than in kl ϩ/ϩ mice (Fig. 1A). Similar results were obtained with five other kl Ϫ/Ϫ mice. The 280-kDa protein band was subjected to in-gel lysylendopeptidase digestion (22), and the sequences of two of the resulting peptides were determined to be LQTASDESYK and KHEAFETDFTVHK by a combination of Edman degradation and mass spectrometry. A data base search of protein sequences revealed that these peptide sequences were homologous to those of human ␣ II -spectrin (GenBank TM accession number AAB41498).
A Western blot using an anti-␣ II -spectrin antibody (C-20) confirmed that the 280-kDa protein is ␣ II -spectrin and that the reactivity of the antibody was drastically decreased in kl Ϫ/Ϫ mice (Fig. 1B). The antibody also stained a 145-kDa band in kl Ϫ/Ϫ mice, but this band was below the detectable level in kl ϩ/ϩ mice (Fig. 1B). Because the anti-␣ II -spectrin antibody recognizes the C terminus of ␣ II -spectrin, it is likely that the 145-kDa band is a C-terminal fragment of ␣ II -spectrin. Although the kidney of 4-week-old kl Ϫ/Ϫ mice was not morphologically different from that of kl ϩ/ϩ mice, it did show a small amount of calcification ( Fig. 1, C-F).
Decrement of ␣ II -Spectrin and Calpastatin with Increased Activation of -Calpain-␣ II -Spectrin was previously shown to be cleaved at a particular site by calpain (29), yielding 136-and 148-kDa fragments. To determine whether calpain is involved in proteolysis of ␣ II -spectrin in the kidney of kl Ϫ/Ϫ mice, we prepared specific antibodies to sequences on either side of the cleavage site (Fig. 2). The anti-BDP-136 antibody, which was produced against a sequence (QQQEVY) in the C-terminal region of BDP-136, recognized only the 136-kDa fragment of ␣ II -spectrin. BDP-136 was detected only in kl Ϫ/Ϫ mice (Fig. 3A). On the other hand, the anti-BDP-148 antibody, which was produced against a sequence (GAMPRD) in the N-terminal region of BDP-148, recognized not only the 148 kDa fragment but also full-length ␣ II -spectrin. BDP-148 was detected only in kl Ϫ/Ϫ mice (Fig. 3B). These results indicated that ␣II-spectrin was degraded by calpain in the kidney of kl Ϫ/Ϫ mice.
To determine which of the calpain isozymes were activated in the kl Ϫ/Ϫ kidney, we made a Western blot of kidney cytosolic fractions using antibodies against four types of calpain: the inactive and active forms of -calpain (pre-and post--calpain) and the inactive and active forms of m-calpain (pre-and postm-calpain). Pre--calpain was detected in kl ϩ/ϩ mice but not in kl Ϫ/Ϫ mice (Fig. 3C). Post--calpain was detected in kl Ϫ/Ϫ mice but not in kl ϩ/ϩ mice (Fig. 3D). Pre-m-calpain was detected in both kl ϩ/ϩ and kl Ϫ/Ϫ mice with no significant difference between them (Fig. 3E). Post-m-calpain was barely detected in either kl ϩ/ϩ or kl Ϫ/Ϫ mice (Fig. 3F). These results indicate that -calpain, but not m-calpain, was specifically activated in the kl Ϫ/Ϫ kidney. Interestingly, calpastatin, which is an endogenous inhibitor of calpain, was barely detected in kl Ϫ/Ϫ mice (Fig. 3G). The triplet bands at about 122 kDa in Fig. 3G are probably alternative splicing forms of calpastatin (30).
The expression levels of mRNAs of calpastatin (Fig. 3H) and ␣ II -spectrin (Fig. 3I) were not different between kl Ϫ/Ϫ and kl ϩ/ϩ mice, which suggests that the decreases of calpastatin and ␣ II -spectrin in kl Ϫ/Ϫ mice were due to increased degradation rather than a down-regulation of transcription.
Activation of -Calpain Depends on Klotho Protein Level-To elucidate the relation between the amount of Klotho protein and the degree of -calpain activation, we examined the mouse heterozygotes for the klotho mutation (kl ϩ/Ϫ ). The expression level of Klotho protein in 2-week-old kl ϩ/Ϫ mice (Fig. 4A, lane  2), was approximately half that in 2-week-old kl ϩ/ϩ mice (lane 1). A similar relation was found in 3-week-old mice (lanes 5 and  5, 4, and 6, respectively). These results showed that the expression level of Klotho protein affected the activation of -calpain and the amount of calpastatin (Fig. 4B).
To elucidate the process of calpain activation, calpastatin decrement, and ␣ II -spectrin proteolysis, we examined mice that were less than 4 weeks old. In kl Ϫ/Ϫ mice, pre--calpain and calpastatin were present at low levels at 2 weeks (Fig. 4A, lane 3) but were undetectable at 3 weeks (lane 6), while the amount of cleaved ␣ II -spectrin was much higher at 3 weeks (lane 6) and 4 weeks than at 2 weeks (lane 3). In 2-and 3-week-old kl ϩ/Ϫ mice (lanes 2 and 5), the amount of cleaved ␣ II -spectrin was much higher at 3 weeks (lane 5) than at 2 weeks (lane 2), while the levels of pre--calpain and calpastatin at 3 weeks were slightly less than those at 2 weeks. These findings suggest that: 1) pre--calpain and calpastatin were originally expressed in kl Ϫ/Ϫ mouse kidney and that -calpain was gradually activated and calpastatin was gradually decreased during development, and 2) ␣ II -spectrin was hardly cleaved in the presence of calpastatin, but intensive cleavage of ␣ II -spectrin was observed after the complete disappearance of calpastatin (Fig. 4B). No calcification was observed in 3-week-old kl Ϫ/Ϫ mice (data not shown), indicating that degradation of ␣ II -spectrin in the kidney of kl Ϫ/Ϫ mice preceded the occurrence of any tissue damage.
Organ-specific Calpain Activation-The susceptibility and degree of proteolysis due to the klotho mutation varied among different organs. Changes in the lung of 4-week-old kl Ϫ/Ϫ mice (Fig. 5) were similar to those observed in the kidney. The intensity of intact ␣ II -spectrin drastically decreased and lower molecular weight bands newly appeared. In addition, post-calpain, but not pre--calpain, was detected, suggesting that significant proteolysis occurred in the lung. It may be relevant to that the first observation of the pulmonary emphysematous changes occurs at 4 weeks of age in kl Ϫ/Ϫ mice (31). In the heart, partial activation of calpain was observed at 4 weeks, and only post--calpain was detected at 8 weeks. However, ␣ II -spectrin was not cleaved in the heart. These results suggest that the heart has a sufficient amount of calpastatin to prevent ␣ II -spectrin degradation. On the other hand, no ␣ II -spectrin degradation or calpain activation was observed in the brain or liver at 8 weeks.
Calpain Activation in Aged Normal Mice-Changes similar to those observed in kl Ϫ/Ϫ mice occurred in aged normal (C57BL/6) mice. As normal mice aged from 4 weeks to 29 months, the expression of Klotho protein decreased, the activation of -calpain increased, the level of calpastatin considerably decreased, and the degradation of ␣ II -spectrin increased (Fig. 6). Similar changes were observed in five other mice. DISCUSSION Our results show that the aberrant activation of -calpain and the decrease of calpastatin in the kidney are caused by the klotho mutation, and such changes lead to the cleavage of ␣ II -spectrin. These phenomena are well correlated with the expression level of Klotho protein. Our results also show that similar changes in -calpain, calpastatin, and ␣ II -spectrin occur in normal aged mice. The abnormal activation of -calpain in the kidney occurs at an early age: in kl Ϫ/Ϫ mice, changes in -calpain activation and ␣ II -spectrin degradation started to occur one to 2 weeks before the appearance of abnormal phenotypes, and in kl ϩ/Ϫ mice, -calpain was gradually activated as they aged, even though these mice have a normal phenotypic appearance.
Our finding that -calpain, but not m-calpain, was activated in the kidney of kl Ϫ/Ϫ mice suggests that the concentration of intracellular calcium ions in these mice is in the micromolar range. Normally calpain is activated temporarily and calpaincatalyzed proteolysis leads to modulation rather than destruction of the substrate proteins. Therefore, continuous activation of -calpain is unusual and elucidation of the mechanism is essential to understanding its pathophysiological role. One possible mechanism is that -calpain overactivation causes a deficiency of calpastatin, and another is that a decrease in calpastatin causes an increase in -calpain activation. Since the transcription levels of calpastatin are the same in kl Ϫ/Ϫ and kl ϩ/ϩ mice, the latter possibility is unlikely, but we cannot completely rule it out. -Calpain activity, in addition to being regulated by the calcium ion concentration, is usually also regulated by the binding of calpastatin (16). Thus, a deficiency of calpastatin may induce the overdestruction of substrates such as ␣ II -spectrin by calpain. The mechanism by which Klotho protein might regulate -calpain activity and calpastatin level in the kidney is unknown. However, it is possible that this regulation is mediated by nitric oxide (NO). NO has been shown to inhibit calpain-mediated proteolysis (32), and sys- temic NO synthesis is decreased in kl Ϫ/Ϫ mice (33,34). Furthermore, adenovirus-mediated klotho gene delivery increased NO production and restored vascular endothelial dysfunction (35). It is noteworthy that calpain overactivation in kl Ϫ/Ϫ mice is not caused by ischemia due to arteriosclerosis, while ischemia would cause overactivation of calpain (13). A previous study revealed that, in kl Ϫ/Ϫ mice, arteriosclerosis first appeared around 4 weeks after birth and progressed gradually with age (1). However, in the lung and kidney in kl Ϫ/Ϫ mice, arteriosclerosis could not be the cause of overactivation of -calpain, because the latter occurred as early as 2ϳ3 weeks.
The degree of proteolysis and of activation of calpain caused by the klotho mutation varied among different organs. Both ␣ II -spectrin degradation and calpain activation were observed in the kidney and lung as early as 2ϳ3 weeks. Spectrin was not cleaved in the heart even at 8 weeks, while overactivation of calpain was observed. The time course of activation of calpain in the heart seemed to be proceeded slower than in the lung and kidney. However, it is impossible to examine this possibility, because kl Ϫ/Ϫ mice die at ϳ8 -9 weeks (1). On the other hand, no ␣ II -spectrin degradation or calpain activation was observed in the brain or liver at 8 weeks. In addition, an organ's susceptibility to the klotho mutation did not necessarily correspond to its expression of klotho mRNA. Taken together, these results suggest that Klotho protein or its metabolites may function as a humoral factor. In support of this hypothesis, both mice and human have a secretory form of Klotho protein (28,36,37), and the exogenous klotho gene expressed in the brain and testis could improve systemic aging phenotypes in kl Ϫ/Ϫ mice (1,38). It is important to identify and characterize a target molecule (receptor) that is responsive to Klotho protein or its metabolites. Thus, it may be that the factor most responsible for an organ's sensitivity to the klotho mutation is the density of such a receptor.
Our finding that normal aged mice show changes similar to those in kl Ϫ/Ϫ mice suggests that the decrease of Klotho protein is closely related to aging processes. Recent studies revealed that the expression of klotho gene was gradually reduced in the rat kidney during long term hypertension (39) and that calpastatin was gently degraded also in the kidney of hypertensive rats (40). Furthermore, humans with chronic renal failure commonly develop multiple complications resembling phenotypes observed in kl Ϫ/Ϫ mice (41)(42)(43)(44)(45), and the expression of klotho mRNA and the production of Klotho protein were severely reduced in these patients (46). Taken together, these results suggest that Klotho protein in the kidney protects the progress of age-related renal disorders.
Based on the above results, we propose that tissue deterioration during aging is caused by a decrease of Klotho protein, which leads to a decrease of calpastatin and activation of -calpain, which leads to a degradation of cytoskeletal components such as spectrin. The magnitude of each of these effects correlates with the amount of Klotho expression. A decrease of calpastatin accelerates the activation of -calpain and vice versa. Such deterioration may trigger tissue abnormalities in kl Ϫ/Ϫ mice and aged mice, but Klotho protein may suppress these processes, while the detailed mechanism is not clear yet. Very recently Yoshida et al. (47) reported that calcium and phosphorus homeostasis could be regulated through Klotho function via the action of 1,25-dihydroxyvitamin D due to the impaired regulation of 1␣-hydroxylase gene expression. This deterioration in the vitamin D 3 endocrine system may participate in many of the phenotypes in kl Ϫ/Ϫ mice via toxicity due to increased levels of calcium, phosphorus, and 1,25-dihydroxyvitamin D. It should be noted that when serum concentrations of calcium, phosphorus, and 1,25-dihydroxyvitamin D are re-stored to normal levels, many of phenotypes are improved despite Klotho protein deficiency. 1 Thus, Klotho protein may be a regulator of calcium homeostasis via the vitamin D 3 endocrine system. Alternatively, based on the homology to ␤-glucosidase (1, 2), Klotho protein may function as a glycosidase-like enzyme and modify the glycan moieties of ion channels. Since it is known that glycosylation appears important for the function of ion channels (48 -50), the change of glycosylation may affect calcium homeostasis. In any case, the abnormal activation of calpain due to the decrease of Klotho protein leads to degradation of cytoskeletal elements such as ␣ II -spectrin is likely to be integral to the pathogenic sequence in kl Ϫ/Ϫ mice and recapitulates effects seen in normal aging. Future studies are needed to determine the definitive role of Klotho protein in the regulation of calcium metabolism as well as of intracellular calcium concentration. Such studies will also lead to a better understand of age-related renal abnormalities and to prevent renal diseases in the future.