| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Centre for Diabetes and Endocrine Research (A.A.R., T.S., H.K.C., G.A.M., J.B.P., J.P.W.), School of Medicine, University of Queensland, Princess Alexandra Hospital, Brisbane, Queensland 4102, Australia; and Institute for Molecular Bioscience (A.J.), University of Queensland, Brisbane, Queensland 4072, Australia
Address all correspondence and requests for reprints to: Jonathan P. Whitehead, Centre for Diabetes and Endocrine Research, University of Queensland, Ground Floor, Princess Alexandra Hospital, Ipswich Road, Woolloongabba, Queensland 4102, Australia. E-mail: jwhitehead{at}cder.soms.uq.edu.au.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
A number of agents are now recognized regulators of adiponectin expression and multimer composition. Thiazolidinediones, a class of peroxisome proliferator activated receptor 
agonists commonly used in antidiabetic therapy, increase adipocyte expression of adiponectin [via a peroxisome proliferator activated receptor response element (9)] and simultaneously increase the proportion of circulating HMW multimers (3). Testosterone, although not affecting intracellular production of adiponectin or multimer composition, appears to selectively retard secretion of HMW multimers from adipocytes (10). This is in keeping with the observed sexual dimorphism in circulating adiponectin levels and multimer composition, with males having reduced levels of total and HMW adiponectin (11, 12). What remains to be elucidated are the molecular mechanisms underlying the regulation of adiponectin multimer composition and secretion.
The adiponectin protein consists of three structural domains (Fig. 1A
); the N-terminal variable region (of greatest amino acid sequence divergence between species), the collagenous domain (named so because of its sequence and structural homology with collagen), and the C-terminal globular (receptor-binding/effector) domain (13). Biochemical analyses of adiponectin multimers have revealed the importance of two kinds of molecular interaction that are integral to their structure (see Fig. 1B). First, the protein forms homotrimers via the noncovalent interactions of the collagenous domains in a triple helix. Hydrophobic interactions between globular domains may also contribute to trimer stability and help to initiate triple helix formation (14). Early biochemical studies demonstrated that temperatures over 70 C resulted in breakdown of intratrimer interactions (2). Second, intermolecular disulfide bonds, involving C39 in the variable domain of the murine protein (C36 of human adiponectin), provide the intertrimer links within hexamers and HMW species. Amino acid substitution of this Cys residue produces a protein unable to form multimers larger than trimer, and reduction of disulfide bonds in wild-type (WT) adiponectin similarly collapses hexamer and HMW species to trimer (2, 11, 15). Furthermore, boiling of purified trimers, hexamers, and HMW multimers, resulting in denaturation of each to their composite monomers and dimers, revealed that hexamers and HMW species, unlike trimers, are composed almost exclusively of disulfide-linked dimers, whereas trimers displayed a 1:1 ratio of monomer to dimer (15).
|
2,8-linked disialic acid moiety (Neu5Aca2
8Neu5Aca23Gal) has been reported, although the site(s) and structural and functional significance of this modification have not been determined (16). Wang et al. (17, 18) identified four conserved prolines in the collagenous domain that undergo hydroxylation, and five conserved lysines (one in the variable region and four in the collagenous domain) that are hydroxylated and subsequently glycosylated (see Fig. 1A
). The same investigators showed that conservative amino acid substitution of the collagenous domain lysines in murine adiponectin (17) produced a variant that was no longer glycosylated and was largely unable to enhance insulin action in hepatocytes. In collagen, proline hydroxylation and lysine hydroxylation and glycosylation are known to have an important structural role in triple helix formation, secretion, and stability (19). This, together with the selective requirement of HMW forms for the hepatic effects of adiponectin, prompted us to examine the role of proline hydroxylation and lysine hydroxylation/glycosylation in multimerization of human adiponectin. In the current study, we have employed site-directed mutagenesis and pharmacological approaches in combination with biochemical analyses to assess the role of PTMs, particularly lysine hydroxylation/glycosylation and proline hydroxylation, in adiponectin multimerization and secretion. Our findings indicate an essential requirement for lysine hydroxylation and glycosylation in the formation of HMW multimers, and provide the first evidence that regulation of HMW multimer production by adipocytes may occur via alterations in PTM of adiponectin.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, A and B). Importantly, we observed peptides resulting from differential cleavage at Lys65, Lys68, and Lys77 (Fig. 2B; peptides B, C, D, and E). Trypsin cleavage occurs at C-terminal peptide bonds of lysine and hydroxylysine but not glucosylgalactosyl hydroxy-lysine residues (20). Thus, our fingerprint data provide clear evidence of differential glycosylation of Lys65, Lys68, and Lys77. Lys101 in peptide F was also variably glycosylated (Fig. 2B).
|
). MS/MS sequencing demonstrated that Pro95, situated within the peptide GFPGIQGR, was present in both the unmodified and hydroxylated form (Fig. 2C, peptide H). Residues Pro71 and Pro76 are located within the peptide GDPGLIGPK (Fig. 2C, peptide G). Analysis of the fingerprint data for this peptide revealed masses of 853 Da (base), 869 Da (base + 16 Da), 885 Da (base + 32 Da), and 901 Da (base + 48 Da), and MALDI-TOF-TOF peptide sequencing confirmed the identity of the base peptide. Together, these data suggest that peptide GDPGLIGPK is differentially hydroxylated at Pro71, Pro76, and Lys77.
Competent Expression and Assembly of Human Adiponectin Multimers in CHO and HEK Cells
Waki et al. (2) first characterized the multimer composition of adiponectin from mouse and human sera by SDS-PAGE of non-reduced and non-heat-denatured samples. These analyses clearly demonstrated the existence of multiple HMW species of human adiponectin compared with a single HMW band in mouse serum. Subsequent studies have focused on the murine protein, and have often employed the well-characterized 3T3-L1 murine adipocyte cell line. The lack of an analogous human adipocyte cell line necessitated the creation of a surrogate expression system that would also be easily amenable to genetic manipulation. WT human adiponectin cDNA was cloned from adipose tissue mRNA by RT-PCR and stably transfected into both HEK and Chinese hamster ovary (CHO) cell lines. HEK cells have been used previously for adiponectin production (3, 11), whereas CHO cells were demonstrated to efficiently produce the core-specific lectin (CSL) or mannose-binding protein, a closely related, multimeric plasma protein secreted by hepatocytes (21, 22). Analysis of secreted adiponectin from CHO and HEK stable cell lines revealed comparable multimer profiles (Fig. 3A, lanes 3 and 4). A C-terminally hemagglutinin (HA)-tagged construct was also generated, and the tagged protein was demonstrated to form similar multimers to untagged adiponectin (Fig. 3A, lanes 4 and 5). Comparison of secreted adiponectin from CHO (lane 3) and HEK (lane 4) stable lines with conditioned media from in vitro-differentiated primary human adipocytes (lane 2) and with human serum (lane 1) revealed a similar profile and distribution of multimers, although the hexamer species consistently appeared as a doublet in serum (Fig. 3A). These stable lines were therefore employed for our subsequent investigations.
|
) were generated by site-directed mutagenesis and expressed transiently or stably in CHO cells. Transfected cells were cultured in serum-free medium, and secreted adiponectin was analyzed for multimer composition on gradient gels. Conservative substitution of either the first two (K2–3R) or last two (K4–5R) lysines in the collagenous domain resulted in impaired HMW multimer production and a greater proportion of trimer than was seen in WT adiponectin from transient transfections (Fig. 3B). The phenotype was more severe in K2–3R, suggesting that these two lysines may serve a more critical function than the fourth and fifth lysines. Mutation of all four lysines in the collagenous domain (K2–5R) had an additive effect, resulting in a complete loss of HMW multimers, and this was confirmed by velocity sedimentation on sucrose gradients (Fig. 3, C and D). The disulfide-incompetent mutant, C36S, formed only trimers as has been reported for the corresponding mutant of murine adiponectin (C39S) (2, 11, 15). Comparison of the secretion rates of WT and mutant adiponectin also revealed striking differences, the K2–5R being secreted with approximately 50% reduced efficiency, whereas the C36S was secreted at 1.5–2 times the rate of WT (Fig. 3E). It was therefore possible that the loss of HMW multimers in secreted K2–5R was due to a failure in secretion of these species rather than a failure in their intracellular formation. To address this, lysates from stable cell lines were analyzed for the presence of HMW multimers after 5-h culture in the presence of brefeldin A (BFA) to inhibit secretion and accumulate the protein in the secretory pathway (23, 24, 25). Although an accumulation of HMW multimers was apparent after BFA treatment of WT-expressing cells, K2–5R cell lysates were devoid of HMW multimers in both BFA-treated and untreated cells (Fig. 3F), indicating a defect in formation of these larger multimers by the mutant protein. A decrease in mobility of all species in lysates was routinely observed after culture in BFA and may result from altered processing due to redistribution of Golgi enzymes into the endoplasmic reticulum (ER) with BFA-induced fusion of the two organelles (24, 25). Intriguingly, no intracellular accumulation of trimer was observed under any condition in the stable cell lines or in primary human adipocytes (data not shown) but did occur after overexpression by transient transfection (data not shown). Analysis of the KR1 mutant containing substitution of the single lysine in the variable domain revealed no apparent effect on multimerization compared with WT (Fig. 4). Similarly, no exacerbation of the quadruple mutant (K2–5R) phenotype was observed upon mutation of all five lysines (K1–5R) (Fig. 4). These observations demonstrate an important functional role for the conserved hydroxylated and glycosylated lysine residues, selectively in the collagenous domain of adiponectin, in the intracellular assembly and secretion of HMW multimers.
|
). Similar results were obtained in primary human adipocytes (data not shown). Analysis of medium samples from the same experiment, after reduction and heat denaturation, revealed a partial inhibition of adiponectin secretion by 2,2'-dipyridyl (Fig. 5A), in agreement with the previously observed effects of the drug on CSL secretion (26). By comparison, BFA completely inhibited secretion (Fig. 5A). Multimer analysis of medium after a longer (24 h) period of secretion revealed that adiponectin secreted from 2,2'-dipyridyl-treated cells is predominantly in the form of trimer (Fig. 5B). Note that multimers in 2,2'-dipyridyl-treated cells migrate more rapidly than those from untreated and BFA-treated cells, consistent with a loss of PTMs with 2,2'-dipyridyl (Fig. 5B). Thus, inhibition of proline hydroxylation in addition to lysine hydroxylation resulted in a more severe impairment of multimer formation, with loss of both hexamer and HMW species. The observed inhibition of adiponectin secretion by 2,2'-dipyridyl mirrors that observed for the K2–5R mutant and is likely due to the loss of hydroxyproline and glucosylgalactosyl hydroxy-lysine residues. However, we cannot exclude the possibility that nonspecific/indirect effects of the drug may also compromise adiponectin multimerization and secretion. Indeed, fragmentation of the Golgi apparatus was observed by immunofluorescence labeling for the cis/medial Golgi marker, GM130, and may well be accompanied by a widespread disruption to ER-to-Golgi transport and/or exocytic vesicle traffic (Fig. 5C). However, dispersion of the Golgi marker was also apparent in BFA-treated cells (Fig. 5C), where formation of HMW adiponectin multimers was uncompromised.
|
). In all analyses, the bacterially produced protein exhibited significantly greater mobility than CHO-Ad, consistent with a lack of eukaryotic PTMs (Fig. 6A, compare lanes 4 and 6). When samples were boiled to disassemble trimers and larger multimers to their constituent monomers and dimers (2), CHO-Ad was present almost exclusively as dimer. In contrast, bacterially produced adiponectin was largely monomeric with only a trace of dimer (Fig. 6A). Analysis of the same samples after both reduction and boiling revealed complete conversion of the CHO-Ad dimer to monomer as well as loss of the bacterial adiponectin dimer species (Fig. 6A). Thus, disulfide bond formation in the bacterially expressed protein appears to be inefficient, at least in the absence of periplasmic targeting. Furthermore, upon analysis of multimer composition on gradient gels, trimer, hexamer, and HMW species were obviously lacking in the bacterially expressed protein, despite the presence of some minor bands in the 60- to 250-kDa size range (Fig. 6B). These species, although disulfide dependent, failed to display the biochemical characteristics of adiponectin multimers from mammalian cells (CHO-Ad), because they were insensitive to boiling (Fig. 6B). Instead, they may represent chaperone-bound intermediates or aggregates. The failure of bacterially expressed adiponectin to multimerize is consistent with the importance of PTMs in this process and has notable implications for functional studies using recombinant adiponectin expressed in prokaryotic systems.
|
). At least as many isoforms were apparent in the C36S and K2–5R mutants, as well as in WT CHO-Ad secreted from dipyridyl-treated cells. (Fig. 7). While WT and C36S samples appeared as a single train of spots, a second train of spots was prominent in both K2–5R mutant and WT CHO-Ad upon dipyridyl treatment. This was accompanied by an increase in the more basic isoforms after dipyridyl treatment, possibly resulting from the loss of acidic hydroxyproline residues. As expected, the loss of neutral sugars in the K2–5R mutant did not effect a dramatic change in pI of adiponectin isoforms. Moreover, the existence of this mutant as a full complement of pI variants indicates that other modifications contributing to pI can still occur in the absence of these glycosylation events. Human adiponectin expressed in bacteria was present as a single isoform with a pI similar to the most basic of the WT CHO-Ad isoforms, consistent with a lack of PTMs.
|
, A and B). To determine whether a change in PTM of the protein may be involved in this response to high glucose, we analyzed the samples by 2DE followed by Western blotting. Adiponectin secreted from the glucose-responsive tissue explants displayed a clearly different pattern of isoforms under high glucose conditions compared with low glucose conditions, whereas no such change was apparent in samples from nonresponsive tissue explants also from female individuals (Fig. 8C). These observations indicate a role for differential PTM of adiponectin in the regulation of its multimer composition by environmental factors during production by adipose tissue under physiologically relevant conditions.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
We have also attempted to address the function of hydroxyprolines in the collagenous domain of human adiponectin with the use of an inhibitor of both lysine and proline hydroxylation, 2',2-dipyridyl. Although we observed an additive effect of this drug over that of mutating the five modified lysine residues, interpretation is made difficult by the clear disruption of the Golgi compartment in treated cells. It is notable, however, that BFA treatment, which completely inhibits secretion and causes fusion of the Golgi compartment with the ER, did not inhibit multimerization of adiponectin.
As a third approach, we expressed human adiponectin in bacteria to produce a minimally modified protein lacking proline and lysine hydroxylation as well as any other eukaryotic PTMs. 2DE analysis revealed a single isoform for the bacterially expressed protein, confirming the loss of differential PTMs. By comparison, adiponectin produced by CHO cells consisted of seven or more pI variants of increasing acidity and molecular weight. The lack of multimerization of the bacterially expressed protein was not immediately apparent, however, because disulfide-dependent HMW species were observed by non-reducing, non-heat-denaturing SDS-PAGE. However, the resistance of these species to heat denaturation, while CHO-expressed adiponectin multimers were readily susceptible, is indicative of aberrant intermolecular interactions unlike those that characterize bona fide multimers (2). Thus, we conclude that these larger species occur as artifacts of bacterial expression. Early characterizations of bacterially expressed adiponectin demonstrated apparent trimer and hexamer (15, 27) and even HMW species (28), either by gel filtration (15, 27) or SDS-PAGE after chemical cross-linking (27, 28, 29). Tsao et al. (15) also determined that the apparent hexamer species was disulfide dependent by reduction with dithiothreitol (DTT). However, more extensive ultrastructural or biochemical characterization of these species was not performed. The failure of bacterially expressed adiponectin to form even bona fide trimers has important implications for functional studies that have employed bacterially produced recombinant adiponectin and provides an explanation for the discrepancies in outcomes from these studies compared with those employing mammalian expression systems. Our observations that the K2–5R mutant, and bacterially expressed adiponectin, fail to form HMW multimers provides mechanistic insight into the inability of this mutant to regulate hepatic glucose output (17), because various lines of evidence indicate that the HMW multimers mediate adiponectin’s effects in the liver (3, 4, 5, 6).
Previous MS analyses of adiponectin have been restricted to the mouse and bovine proteins. To confirm the existence of the same modifications in human adiponectin, we have analyzed tryptic digests of recombinant HEK-produced human adiponectin using this technique. Our data show that human adiponectin is hydroxylated or glycosylated on the same Pro and Lys residues shown to be modified in murine and bovine adiponectin. In addition, we have now identified three novel sites of Pro hydroxylation (on residues 71, 76, and 95) and demonstrate the differential hydroxylation of these in human adiponectin. Amino acid sequence alignments revealed that Pro76 and Pro95 are completely conserved in mammals, whereas Pro71 is restricted to primate and chicken sequences (Fig. 9). Because the laser ionization energy used in the MALDI analysis is far lower than that required for removal of a hydroxyl group from a modified lysine or proline residue, it may be concluded that the variable modification of these residues is not an artifact of mass spectrometric analysis and hence may contribute to the multiple pI isoforms observed by 2DE. The presence of additional differential modifications, such as sialylation (16), may also contribute to the multiple pI isoforms and warrants further investigation.
|
Regulation of adiponectin multimer composition is likely to occur at multiple levels through multiple mechanisms. The molecular mechanisms behind the increase in HMW adiponectin production in response to high glucose remain to be elucidated and will involve identification of the modifications that contribute to the observed changes in pI and mobility. Whether these modifications affect efficiency of HMW formation within the secretory pathway or efficiency of HMW secretion will also be of interest. Xu et al. (10) recently demonstrated selective inhibition of HMW multimer secretion (but not formation) by testosterone in 3T3-L1 adipocytes. This indicates the existence of mechanisms in adipocytes for the selective regulation of HMW multimer secretion.
Our parallel analyses of multimer composition by separation on gradient gels as well as by velocity sedimentation on sucrose gradients (see Figs. 3, C and D, and 8) have provided valuable insights into the relative merits of the two methods. SDS-PAGE under non-reducing and non-heat-denaturing conditions allowed resolution of individual multimers for qualitative analyses, whereas sucrose density gradients permitted only crude fractionation into low-molecular-weight (LMW) (trimer and hexamer) and HMW peaks. This was clearly illustrated by analysis, using both methods, of the K2–5R mutant (Fig. 3), which forms trimers and hexamers but no HMW multimers. Quantitation of the absolute amounts of adiponectin present as LMW and HMW species, on the other hand, requires that multimers be first fractionated according to size (either on sucrose gradients or by size-exclusion chromatography) before reduction and denaturation of fractions to constituent monomers for quantitation by SDS-PAGE and Western blotting.
The 2DE pattern of human adiponectin isoforms secreted from CHO cells bears close resemblance to that described nearly two decades ago for the CSL, a 26-kDa plasma protein secreted by hepatocytes (32). CSL is an important component of innate immunity, binding to the surface carbohydrates of pathogens and mediating complement activation (33). A member of the C1q superfamily of structurally homologous proteins, CSL also undergoes hydroxylation of prolines and lysines and subsequent glycosylation of hydroxylysines (21) and forms intramolecular and intermolecular disulfide bonds (26). Also, like adiponectin, CSL is produced as LMW and HMW multimers that fractionate similarly on 5–20% sucrose gradients (32). The functional significance of CSL multimers has recently been demonstrated by Larsen et al. (22), who analyzed naturally occurring mutants of the protein expressed in CHO cells. Patients having these mutant alleles have reduced circulating CSL and demonstrate immune deficiencies and a greater susceptibility to develop infections (reviewed in Ref. 22). Mechanistic insights were provided by characterization of the secreted mutant proteins, which revealed a deficiency in HMW multimers of CSL that was accompanied by a loss of ligand-binding activity and ability to activate complement. Thus, a similar relationship between HMW multimer assembly and functional activity appears to exist in this adiponectin-related protein.
In conclusion, the present study elucidates a role for lysine hydroxylation and glycosylation in adiponectin assembly into HMW multimers and, furthermore, demonstrates that the glucose-stimulated increase in HMW multimer production by adipose tissue is associated with changes in PTM of adiponectin.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Molecular Cloning and Site-Directed Mutagenesis
WT adiponectin cDNA was generated from human sc adipose tissue mRNA and amplified by reverse transcription and PCR using the following primers: forward, 5'-ATGCTGTTGCTGGGAGC-3'; reverse, 5'- TCAGTTGGTGTCATGG-3'. The amplicon was cloned into the pGEM-T-Easy shuttle vector (Promega), and a 5' Kozak sequence was added (forward primer, 5'-GCCACCATGCTGTTGCTGGGAGC-3') before sequencing and subcloning into the pcDNA5/FRT expression plasmid. Site-directed mutagenesis was performed using the QuikChange site-directed mutagenesis kit from Stratagene (La Jolla, CA) using complementary primers designed according to the manufacturer’s instructions. The following forward primers were used: C36S (5'-cccaagggggccagcacaggttggatgg-3'); K1R (5'-cccctgcccaggggggcctgcacagg-3'); K2–3R (5'-cctggtgagcggggtgagcgaggagatcc-3'); K4R (5'-cttattggtcctcggggagacatcgg-3'); K5R (5'-ccaaggcaggcgaggagaacctgg-3'). Reverse primers were exactly complementary to forward primers. The K2–5R and K1–5R mutants were produced by multiple rounds of mutagenesis, and all mutants were confirmed by sequencing.
CHO Cell Culture, Transfection, and Generation of Stable Cell Lines
CHO cells were cultured in HAM’s F12 medium supplemented with 10% fetal bovine serum and 2 mM L-glutamine and transfected using Lipofectamine Plus reagent according to the manufacturer’s instructions. CHO stable cell lines expressing WT or mutant human adiponectin constructs were generated using the Flp-In system from Invitrogen according to the manufacturer’s protocols. Briefly, WT and mutant adiponectin cDNA in the expression plasmid, pcDNA5/FRT, were cotransfected with the Flp recombinase-encoding plasmid (pOG44) into the Flp-In CHO host cell line (Invitrogen). Forty-eight hours after transfection, cells were reseeded and cultured in Hygromycin (600 µg/ml)-containing medium for 2 wk to select adiponectin-expressing cells. Immunofluorescence microscopy revealed that approximately 50% of cells in the selected population were expressing adiponectin after this time. Note that, although these were not a single clonal population of stables, all stable transfectants contain one copy of the cDNA integrated into a single FRT site at a defined genomic locus of the Flp-In CHO host cell line.
Adipose Tissue Explant Experiments
Subcutaneous adipose tissue samples were obtained from four male [average age, 33.5 (range, 24–51); average BMI, 55 (range, 46.4–67.9)] and 11 female [average age, 41.4 (range, 21–66); average BMI, 45.5 (range, 26–58)] patients undergoing elective open-abdominal surgical procedures. None of the patients had diabetes or severe systemic illness, and none was taking medications known to affect adipose tissue mass or metabolism. The protocol was approved by the research ethics committees of the University of Queensland, Princess Alexandra Hospital, and Wesley Hospital. All patients gave their informed written consent. The tissue was cut into 10- to 30-mg explants, washed, and cultured in 24-well plates (one explant per well) in 300 µl of DMEM (containing 5 mM glucose) overnight. The next morning, the culture medium was removed, and the explants were cultured in fresh DMEM containing 5 or 15 mM glucose for a further 72 h. At harvest, medium samples were cleared to remove cell/tissue debris before analysis on gradient gels, sucrose density gradients, or by 2DE and Western blotting.
Primary Human Adipocyte Culture
Primary human adipocytes, isolated from sc adipose tissue and differentiated in vitro as previously described (34), were kindly provided by Dr. Louise Hutley and Ms. Felicity Newell (Centre for Diabetes and Endocrine Research, Brisbane, Australia). The cells were cultured in DMEM supplemented with 33 µM biotin, 17 µM pantothenate, 10 µg/ml transferrin, 2 mM L-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin during the course of experiments, which were performed on d 21 of differentiation.
Treatment of Cultured Cells/Tissue Explants
For experiments assessing the effects of 5 or 15 mM glucose, cells were cultured in low-glucose DMEM (containing 1000 mg/liter glucose) or a 1:1 mixture of low-glucose DMEM and high-glucose DMEM (containing 4500 mg/liter glucose), respectively. Cycloheximide treatment of adipocytes and stable CHO cell lines was for 20 h using 10 and 50 µg/ml concentrations of the drug for the two cell types, respectively. BFA and 2,2'-dipyridyl were used on both cell types at 5 µg/ml and 3.5 mM, respectively.
Harvesting and Preparation of Lysate and Medium Samples for Analysis
CHO cells were harvested on ice into lysis buffer [50 mM HEPES, 150 mM NaCl, 1% Triton-X-100, 1 mM Na3VO4, 30 mM NaF, 10 mM Na4P2O7, 10 mM EDTA, 1 µg/ml antipain, 1 µg/ml pepstatin, 1 µg/ml leupeptin, 250 µM benzamidine, 400 µM 4-(2-aminoethyl)benzenesulfonyl fluoride] and cleared (16,000 x g for 20 min at 4 C). Protein concentration of lysates was determined using a bicinchoninic acid protein assay (Pierce) and equal protein loaded for analysis by SDS-PAGE. Medium samples were cleared at 16,000 x g for 10 min to remove cell debris and then concentrated by precipitation with trichloroacetic acid or acetone for analysis under reducing and heat-denaturing conditions. For multimer analysis on gradient gels or by velocity sedimentation, samples were concentrated using centricons of 30,000 MW cutoff (Millipore, Billerica, MA) or by ammonium sulfate precipitation (40% wt/vol).
Bacterial Expression of Adiponectin
Human adiponectin cDNA lacking the 18-amino acid mammalian signal sequence was generated by PCR (forward primer, 5'-catatggaaaccacgactcaagggcccgg-3'; reverse primer, 5'-ggatcctcagttggtgtcatggtagagaagaaagcc-3') and cloned into the pET 20b expression plasmid (using NdeI and BamHI restriction sites incorporated into the primers). The insert was sequenced and transformed into BL21(DE3)pLys-S cells for inducible expression. Expression cultures (10 ml) were inoculated from an overnight starter culture and allowed to reach an A600 OD of 0.6 AU before induction with 1 mM IPTG for 3 h at 37 C. Cells were extracted with BugBuster Master Mix lysis buffer (Novagen, Madison, WI) and fractionated into soluble and particulate fractions by centrifugation of lysates at 16,000 x g for 20 min at 4 C. Samples were analyzed by SDS-PAGE and Western blotting.
Velocity Sedimentation on Sucrose Gradients
Fractionation of adiponectin multimers into LMW and HMW multimers was achieved by separation on sucrose gradients essentially as described by Pajvani et al. (3). Briefly, 2-ml gradients of 5–20% sucrose were poured stepwise into thin-walled ultracentrifuge tubes and allowed to equilibrate overnight. A 200-µl volume of sample [diluted as required into sucrose gradient buffer; 10 mM HEPES (pH 8), 125 mM NaCl] was applied to the top of the gradient before centrifugation at 55,000 rpm in a Sorvall S-55 rotor for 4 h at 4 C. Gradients were then fractionated from the top into 15 x 150-µl fractions, which were precipitated with 2 vol of acetone. Fractions were analyzed by SDS-PAGE and Western blotting of equal volumes after reduction and boiling.
SDS-PAGE and Western Blotting
Samples were prepared for SDS-PAGE as described by Waki et al. (2). For analysis under non-reducing, non-heat-denaturing conditions, 5x SDS-PAGE sample buffer lacking reducing agents was added to a final concentration of 1x [3% sodium dodecyl sulfate, 50 mM Tris-HCl (pH 6.8), 10% glycerol], and samples were incubated for 30 min at room temperature (RT) before analysis on 4–12% or 3–10% gradient gels. For analysis of fully denatured, reduced protein, samples were prepared using 5x SDS-PAGE sample buffer containing 500 mM DTT and boiled for 10 min before electrophoresis on 12% acrylamide gels. After electrophoretic separation, proteins were blotted onto polyvinylidene difluoride membranes. The membranes were blocked with 1% fish skin gelatin in PBS for 30 min and probed with a 1:5000 dilution of polyclonal antibody to human adiponectin (Chemicon). Membranes were then washed in PBS containing 0.1% Tween and probed with a fluorescent antirabbit secondary (Rocklands Immunochemicals) before scanning of membranes and quantitation of Westerns using the Odyssey infrared imaging system (LICOR Biosciences, Lincoln, NE). Alternatively, membranes were probed with horseradish peroxidase-conjugated secondary antibody and visualized by chemiluminescence using the Supersignal West Pico Chemiluminescent substrate system from Pierce.
2DE
Samples were analyzed by 2DE using the Zoom 2DE system and associated reagents from Invitrogen according to the manufacturer’s protocols. In brief, samples of conditioned medium were precipitated with acetone, and the pellets were resuspended in lysis buffer containing ZOOM 2DE Protein Solubilizer 2 and 20 mM DTT. After a 15-min incubation at RT, samples were alkylated by addition of 0.5% N,N-dimethylacrylamide and incubation for 30 min. Excess DMA was neutralized by further addition of DTT to 40 mM followed by a clearing spin (16,000 x g, 20 min, 4 C) and acetone precipitation. Protein pellets were resuspended in rehydration buffer containing ZOOM 2DE Protein Solubilizer 2, 10 mM DTT, and 0.5% carrier ampholytes (pH 4–7), and applied to ZOOM strips (pH 4–7) for overnight rehydration. Isoelectric focusing was performed using a ZOOM IPG Runner System followed by reduction and alkylation of the strips according to the manufacturer’s instructions. Second-phase separation was carried out using 12% minigels and followed by Western blotting.
MALDI-TOF, MALDI/TOF/TOF, and MS/MS analyses
Adiponectin (2 µg) was subject to overnight digestion with trypsin (100 ng) by coprecipitation in methanol at –20 C, resuspension in 10 µl of 40 mM ammonium bicarbonate/10% acetonitrile, and overnight incubation at 37 C. A trypsin-only digest (without adiponectin) was performed as negative control. A virtual trypsin digest of adiponectin (to obtain predicted peptide masses) was performed using MS-Digest software on the Protein Prospector proteomics server.
MALDI-TOF analysis was performed on an Applied Biosystems (ABI) Voyager-DE STR Biospectrometry Workstation. A total of 0.5 µl of the digest was mixed with an equal volume of -cyano-4-hydroxy-cinnamic acid matrix solution (5 mg/ml in EtOH), spotted, and allowed to dry on the sample plate. All MALDI-TOF analysis was performed in positive reflector mode using a pulsed laser and 20,000-V acceleration. Spectra were mass calibrated using an internal standard mix of angiotensin II (MH+ 1046.5), human ACTH (18–39 clip MH+ 2465.2), and oxidized insulin B chain (MH+ 3494.6). Spectra were analyzed for expected peptide masses using the ABI Data Explorer software. Peptide candidates with probable carbohydrate and hydroxyl group modifications were identified by comparing the experimental and theoretical mass fingerprints. These candidate peptides were submitted for MS/MS and MALDI-TOF-TOF tandem MS sequencing to confirm the amino acid sequence and identify modified residues.
LC/MS/MS sequencing experiments were performed on an ABI 4000 Q-TRAP LC/MS/MS System. Before LC/MS/MS analysis, the tryptic peptide sample was diluted in 40 µl of a 1% trifluoroacetic acid/1% formic acid (AQ) solution and separated on a 0.5-mm (inner diameter) by 150-mm (length) Luna 5µ C18 RP-HPLC (Phenomenex, Torrance, CA) column using a flow rate of 30 µl/min. The eluted peptides were directly applied to the 4000 Q-TRAP MS/MS workstation. Peptide fragmentation data were collected and analyzed using the ABI Analyst QS software package, and sequences were scored using the MASCOT search database.
MALDI-TOF-TOF sequencing experiments were performed on an ABI 4700 Proteomics Discovery System. A total of 0.4 µl of the digest was mixed with an equal volume of -cyano-4-hydroxy-cinnamic acid matrix solution (5 mg/ml in 60/40 methyl cyanide /0.1% formic acid in 5 mM ammonium phosphate), spotted, and allowed to dry on the sample plate. MS and MS/MS experiments were carried out using 1-kV MS and MS/MS operating modes and 200-Hz laser repetition rate. MS data was externally calibrated using the ABI calibration mix plus Angiotensin (MH+ 1296.685). MS/MS mass calibration was carried out using the fragments from the collision-induced dissociation of the external standard Angiotensin. Precursor ions with predicted modifications from MALDI-TOF MASS fingerprints were collision cell fragmented and sequenced. Sequence analysis was performed using the ABI GPS and DeNovo Explorer software and ranked using the MASCOT search database.
Immunofluorescence Microscopy
Cells were grown on 12-mm round glass coverslips and processed as per Richards et al. (35). Briefly, cells were fixed for 20 min in 4% paraformaldehyde and permeabilized in ice-cold methanol for 7 min at –20 C. A 10-min incubation in 50 mM NH4Cl at RT was followed by a 20-min blocking step in a solution of 0.2% fish skin gelatin and 0.2% BSA in PBS. Cells were incubated in primary antibody solution for 30 min at RT before four 5-min PBS washes and a 20-min incubation in fluorophore-conjugated secondary antibody. After another four 5-min PBS washes, cells were rinsed in MilliQ water and mounted in Moviol mounting medium (Merck Biosciences, San Diego, CA). Specimens were analyzed on an epifluorescence microscope, and images were prepared using Photoshop 7.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
A.A.R., T.S., H.K.C., A.J., G.A.M., J.B.P., and J.P.W. have nothing to declare.
First Published Online February 23, 2006
Abbreviations: BFA, Brefeldin A; BMI, body mass index; CHO, Chinese hamster ovary; CSL, core-specific lectin; 2DE, two-dimensional electrophoresis; DTT, dithiothreitol; ER, endoplasmic reticulum; HA, hemagglutinin; HEK, human embryonic kidney; HMW, high molecular weight; IPTG, isopropyl-ß-D-thiogalactopyranoside; LC, liquid chromatography; LMW, low molecular weight; MALDI-TOF, matrix-assisted laser desorption/ionization-time of flight; MS, mass spectrometry; PTM, posttranslational modification; RT, room temperature; WT, wild type.
Received for publication September 22, 2005. Accepted for publication February 10, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
B signaling pathway by adipocyte complement-related protein of 30 kDa (Acrp30). J Biol Chem 277:29359–29362This article has been cited by other articles:
 |
|
 |
|
  G. L. Hendricks III, J. A. Hadley, S. M. Krzysik-Walker, K. S. Prabhu, R. Vasilatos-Younken, and R. Ramachandran Unique Profile of Chicken Adiponectin, a Predominantly Heavy Molecular Weight Multimer, and Relationship to Visceral Adiposity Endocrinology, July 1, 2009; 150(7): 3092 - 3100. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Banga, R. Unal, P. Tripathi, I. Pokrovskaya, R. J. Owens, P. A. Kern, and G. Ranganathan Adiponectin translation is increased by the PPAR{gamma} agonists pioglitazone and {omega}-3 fatty acids Am J Physiol Endocrinol Metab, March 1, 2009; 296(3): E480 - E489. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Schegg, A. J. Hulsmeier, C. Rutschmann, C. Maag, and T. Hennet Core Glycosylation of Collagen Is Initiated by Two {beta}(1-O)Galactosyltransferases Mol. Cell. Biol., February 15, 2009; 29(4): 943 - 952. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. W. Wong, S. A. Krawczyk, C. Kitidis-Mitrokostas, G. Ge, E. Spooner, C. Hug, R. Gimeno, and H. F. Lodish Identification and characterization of CTRP9, a novel secreted glycoprotein, from adipose tissue that reduces serum glucose in mice and forms heterotrimers with adiponectin FASEB J, January 1, 2009; 23(1): 241 - 258. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Teoh, A. Quan, K. W. A. Bang, G. Wang, F. Lovren, V. Vu, J. J. Haitsma, P. E. Szmitko, M. Al-Omran, C.-H. Wang, et al. Adiponectin deficiency promotes endothelial activation and profoundly exacerbates sepsis-related mortality Am J Physiol Endocrinol Metab, September 1, 2008; 295(3): E658 - E664. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Leth, K. K. Andersen, J. Frystyk, L. Tarnow, P. Rossing, H.-H. Parving, and A. Flyvbjerg Elevated Levels of High-Molecular-Weight Adiponectin in Type 1 Diabetes J. Clin. Endocrinol. Metab., August 1, 2008; 93(8): 3186 - 3191. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Saraheimo, C. Forsblom, L. Thorn, J. Waden, M. Rosengard-Barlund, O. Heikkila, K. Hietala, D. Gordin, J. Frystyk, A. Flyvbjerg, et al. Serum Adiponectin and Progression of Diabetic Nephropathy in Patients With Type 1 Diabetes Diabetes Care, June 1, 2008; 31(6): 1165 - 1169. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Miller, P. Siripurkpong, J. Hawes, A. Majdalawieh, H.-S. Ro, and R. S. McLeod The trans-10, cis-12 isomer of conjugated linoleic acid decreases adiponectin assembly by PPAR{gamma}-dependent and PPAR{gamma}-independent mechanisms J. Lipid Res., March 1, 2008; 49(3): 550 - 562. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Chandrasekar, D. N. Patel, S. Mummidi, J.-w. Kim, R. A. Clark, and A. J. Valente Interleukin-18 Suppresses Adiponectin Expression in 3T3-L1 Adipocytes via a Novel Signal Transduction Pathway Involving ERK1/2-dependent NFATc4 Phosphorylation J. Biol. Chem., February 15, 2008; 283(7): 4200 - 4209. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Sipila, H. Ruotsalainen, R. Sormunen, N. L. Baker, S. R. Lamande, M. Vapola, C. Wang, Y. Sado, A. Aszodi, and R. Myllyla Secretion and Assembly of Type IV and VI Collagens Depend on Glycosylation of Hydroxylysines J. Biol. Chem., November 16, 2007; 282(46): 33381 - 33388. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Maahs, L. G. Ogden, J. K. Snell-Bergeon, G. L. Kinney, R. P. Wadwa, J. E. Hokanson, D. Dabelea, A. Kretowski, R. H. Eckel, and M. Rewers Determinants of Serum Adiponectin in Persons with and without Type 1 Diabetes Am. J. Epidemiol., September 15, 2007; 166(6): 731 - 740. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. W Peake, J. T Hughes, Y. Shen, and J. A Charlesworth Glycosylation of human adiponectin affects its conformation and stability J. Mol. Endocrinol., July 1, 2007; 39(1): 45 - 52. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. J Simons, P. S van den Pangaart, J. M F G Aerts, and L. Boon Pro-inflammatory delipidizing cytokines reduce adiponectin secretion from human adipocytes without affecting adiponectin oligomerization J. Endocrinol., February 1, 2007; 192(2): 289 - 299. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Bodles, A. Banga, N. Rasouli, F. Ono, P. A. Kern, and R. J. Owens Pioglitazone increases secretion of high-molecular-weight adiponectin from adipocytes Am J Physiol Endocrinol Metab, November 1, 2006; 291(5): E1100 - E1105. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
