1H NMR-based metabolomic study on resistance to diet-induced obesity in AHNAK knock-out mice

AHNAK is a giant protein of approximately 700 kDa identified in human neuroblastomas and skin epithelial cells. Recently, we found that AHNAK knock-out (AHNAK −/− ) mice have a strong resistance to high-fat diet-induced obesity. In this study, we applied 1 H NMR-based metabolomics with multivariate statistical analysis to compare the altered metabolic patterns detected in urine from high-fat diet (HFD) fed wild-type and AHNAK −/− mice and investigate the mechanisms underlying the resistance to high-fat diet-induced obesity in AHNAK −/− mice. In global profiling, principal components analysis showed a clear separation between the chow diet and HFD groups; wild-type and AHNAK −/− mice were more distinctly separated in the HFD group compared to the chow diet group. Based on target profiling, the urinary metabolites of HFD-fed AHNAK −/− mice gave higher levels of methionine, putrescine, tartrate, urocanate, sucrose, glucose, threonine, and 3-hydroxyisovalerate. Furthermore, two-way ANOVAs indicated that diet type, genetic type, and their interaction (gene × diet) affect the metabolite changes differently. Most metabolites were affected by diet type, and putrescine, threonine, urocanate, and tartrate were also affected by genetic type. In addition, cis -aconitate, succinate, glycine, histidine, methylamine (MA), phenylacetylglycine (PAG), methionine, putrescine, uroconate, and tartrate showed interaction effects. Through the pattern changes in urinary metabolites of HFD-fed AHNAK −/− mice, our data suggest that the strong resistance to HFD-induced obesity in AHNAK −/− mice comes from perturbations of amino acids, such as methionine, putrescine, threonine, and histidine, which are related to fat metabolism. The changes in metabolites affected by microflora such as PAG and MA were also observed. In addition, resistance to obesity in HFD-fed AHNAK −/− mice was not related to an activated tricarboxylic acid cycle. These findings demonstrate that 1 H NMR-based metabolic profiling of urine is suitable for elucidating possible biological pathways perturbed by functional loss of AHNAK on HFD feeding and could elucidate the mechanism underlying the resistance to high-fat diet-induced obesity in AHNAK −/− mice. Keywords Metabolomics AHNAK High-fat diet Metabolic profiling 1 H NMR 1 Introduction AHNAK is a protein that plays a key regulator function in calcium homeostasis. In epithelial cells, AHNAK is present mainly in the cytoplasm when cells are kept in low Ca 2+ medium, but translocates to the plasma membrane after an increase in extracellular Ca 2+ concentration or protein kinase C activation [1] . In addition, AHNAK has been found in cardiomyocytes associated with L-type calcium channels. In these cells, AHNAK may play a role in cardiac calcium signaling by modulating L-type calcium channels in response to β-adrenergic stimulation [2] . Sekiya et al. [3] also identified AHNAK as a protein that binds and activates phospholipase C-g1 in the presence of arachidonic acid to generate inositol trisphosphate and diacylglycerol, two second messengers that control intracellular calcium flux. Recently, genetically engineered mice are used in various human disease researches to know gene function and their molecular pathogenesis [4] , AHNAK knockout (KO) mice were generated by two different research institutions [5,6] , but the resulting phenotypes or clinical findings are not known. The only published data were related to the underdevelopment of T cells in AHNAK KO mice [7] . However, increasing intracellular calcium via the stimulation of either receptor- or voltage-mediated calcium channels has been shown to stimulate the expression and activity of fatty acid synthesis and a key enzyme in de novo lipogenesis, and inhibit basal and agonist-stimulated lipolysis in both human and murine adipocytes [8–10] . Therefore, AHNAK null (AHNAK −/− ) mice were expected to have a lower body weight compared to the wild-type, and AHNAK KO mice were recently determined to show a strong resistance to diet-induced obesity on a high-fat diet (HFD). However, the mechanism that inhibits lipogenesis in AHNAK −/− mice is unknown. Thus, we investigated the changes in metabolites, which are end products of gene expression, and identified the biochemical phenotype in the urine of AHNAK −/− mice fed the HFD. Metabolomics, the multi-targeted analysis of endogenous metabolites from biological samples, is a recently developed technique that is useful in understanding pathophysiologic processes, detecting the metabolic profiles associated with genetic defects, and discriminating between phenotypes of experimental animals [11,12] . The present study demonstrates the biological processes related to metabolite perturbations and the mechanism underlying the resistance to diet-induced obesity in AHNAK −/− mice using NMR-based metabolomics. 2 Materials and methods 2.1 Animal experiments The generation of AHNAK KO mice has been described previously [13] . Wild-type and AHNAK −/− mice used for comparison were littermates. Mice used for data analysis in this study were male unless otherwise indicated. The animals were maintained at 24 ± 2 °C with a 12-h light/dark cycle and fed a standard chow diet (LabDiet, Richmond, IN, USA) ad libitum with tap water. At 6 weeks of age, wild-type AHNAK −/− and mice fed the HFD (20% carbohydrate, 60% fat, 20% protein: D12492 : Research Diets Inc., New Brunswick, NJ, USA) for 12 weeks. These procedures were reviewed according to the “Guide for Animal Experiments” (edited by the Korean Academy of Medical Sciences) by the Institutional Animal Care and Use Committees (IACUC) of Seoul National University. 2.2 1 H NMR spectroscopic analysis of urine The urine samples were stored at −80 °C. These samples were thawed at room temperature and centrifuged at 13,000 rpm for 10 min prior to NMR analysis. For the NMR analysis, 400 μL of urine supernatant was mixed with 200 μL of phosphate-buffer solution (0.2 M, 0.018% NaN 3 ) and adjusted to pH 7.0 ± 0.1. Aliquots of 540 μL of the supernatant and 60 μL of D 2 O (containing 5 mM sodium 2,2-dimethyl-2-silapentane-5-sulfonate (DSS) for internal chemical standard) were transferred into 5-mm NMR tubes. 1 H NMR spectra were collected on a VNMRS-600 MHz NMR spectrometer (Varian Inc., Palo Alto, CA, USA) using a triple-resonance 5-mm HCN salt-tolerant cold probe. Water was suppressed by the standard one-dimensional (1D) Noesypresat pulse sequence (RD–90°–t1–90°–tm–90°–FID acquisition). For each urine sample, the spectrum was collected with 64 transients into 67,568 data points using a spectral width of 8445.9 Hz, relaxation delay of 2.0 s, and an acquisition time of 4.0 s. 2.3 NMR spectral data reduction and preprocessing All NMR spectra data were segmented into equal widths (0.005 ppm) corresponding to the regions δ0.7–δ9.5 ppm. The region corresponding to water, urea (δ4.5–δ6.2), and DSS (δ1.70–δ1.80, δ2.89–δ2.97) regions were removed prior to normalization and spectra alignment. The spectra were normalized to the total spectral area and converted to ASCII format files. The ASCII files were imported to MATLAB (R2008a, Mathworks, Inc., 2008), and all spectra were aligned using the correlation optimized warping method [14] . The resulting data sets were imported into SIMCA-P version 12.0 (Umetrics, Umeå, Sweden) for multivariate analysis. NMR spectral data analysis was accomplished using targeted profiling with Chenomx NMR Suite 6.0 (Chenomx Inc., Edmonton, AB, Canada) and concentrations were determined using the 600 MHz library from Chenomx NMR Suite 6.0, which compares the integral of a known reference signal (DSS-d6) with signals derived from a library of compounds containing chemical shifts and peak multiplicities. 2.4 Multivariate analysis The resulting data sets were imported into SIMCA-P version 12.0 (Umetrics) for multivariate analysis. All imported data were mean-centered for multivariate data analysis. PCA, an unsupervised pattern recognition method, was performed to examine the intrinsic variation in the data set. The quality of each model was determined by the goodness of fit parameter ( R 2 ) and a goodness of prediction parameter ( Q 2 ) [15] . A two-way ANOVA with Bonferroni post-test was applied to the concentrations of urinary metabolites for gene and diet using Prism version 5 for Windows (GraphPad Software, San Diego, CA, USA). Statistical significance was set at P < 0.05. 3 Results 3.1 Body and fat weight AHNAK −/− mice weighed significantly less compared to wild-type controls ( Fig. 1 A). The AHNAK −/− mice had approximately 20% less body weight than wild-type mice feeding with a chow diet. The lower body weight in AHNAK −/− mice might have been related to the resistance to diet-induced obesity. To test this hypothesis, the wild-type and AHNAK −/− mice were fed a HFD in which 60% of the calories were derived from fat. As expected, the wild-type mice gained significant weight, but the AHNAK −/− mice were unaffected ( Fig. 1 A). The difference in body weight became more pronounced on the HFD than chow diet. In addition, fat weight and the fat to body weight ratio were lower in AHNAK −/− mice. In the HFD groups, AHNAK −/− mice exhibited significantly lower fat weight and fat to body weight ratio than wild-type mice ( Fig. 1 B), indicating that the AHNAK −/− mice are protected from HFD-induced obesity. 3.2 1 H NMR spectroscopy and pattern recognition analysis Fig. 2 shows representative 600-MHz 1D 1 H NMR spectra for urine from wild-type and AHNAK −/− mice on the chow diet ( Fig. 2 A and B), and HFD ( Fig. 2 C and D). Visual inspection of 1 H NMR spectra showed significant differences among all four groups. Spectral resonances of metabolites were assigned according to the 600 MHz library from Chenomx NMR Suite 6.0. NMR spectra of urine were dominated by numerous metabolites as shown in Fig. 2 . Principal component analysis (PCA) was applied to the NMR spectra to investigate the intrinsic differences in the metabolite levels of all groups tested ( Fig. 3 A). The PCA score plot derived from the NMR spectra showed a clear separation between the chow diet group and the HFD group along PC1 ( R 2 = 0.429, Q 2 = 0.246). In the HFD groups, the AHNAK −/− mice showed a distinct separation from the wild-type along PC2. This separation was not apparent in the chow diet groups. These clustering patterns are consistent with the differences in body weight between the wild-type and AHNAK −/− mice in the HFD groups. The big difference in body weight between HFD fed wild-type and AHNAK −/− mice results in the metabolic variation in HFD fed groups than chow diet groups. 3.3 Targeted metabolic profiling Endogenous metabolites were identified and then quantified using the 600 MHz library from the Chenomx NMR suite. Thirty-four metabolites comprising the observed differences between the groups are listed in Table 1 . The differences in the levels of the quantified urinary metabolites between the groups were compared using multivariate analysis. The PCA score plot ( Fig. 3 B) of the urinary metabolite concentrations obtained for all groups showed a diet-dependent separation along PC1 and a genetic-dependent separation along PC2 ( R 2 = 0.797, Q 2 = 0.557). This target profiling result indicated that the discrimination by the diet difference is much more pronounced compared to the genetic difference. To identity the metabolites responsible for the differentiation in the PCA score plots between groups, PCA loading plots were generated ( Fig. 3 C). This plot shows increased levels of methionine, putrescine, tartrate, uroconate, sucrose, glucose, threonine, and 3-hydroxyisovalerate (HIV) in HFD-fed AHNAK −/− mice, whereas no significantly increased metabolites were observed in HFD-fed wild-type mice. 3.4 Metabolomic analysis for the response of the AHNAK gene and high-fat diet To investigate the effect of the diet type (HFD) and gene type (AHNAK −/− ) in metabolite changes, two-way ANOVAs were applied to measure metabolite concentrations ( Table 1 ). Most of the metabolites (25/34) were affected by the HFD, whereas putrescine, threonine, urocanate, and tartrate were affected by gene type. Furthermore, cis -aconitate, succinate, glycine, histidine, methylamine (MA), phenylacetylglycine (PAG), methionine, putrescine, uroconate, and tartrate showed an interaction (gene × diet) effect. However, urocanate and tartrate were only found in the urine of HFD-fed mice ( Fig. 4 ). The levels of MA in wild-type mice were lowered in the HFD group compared to the chow diet group, whereas those of AHNAK −/− mice showed no significant difference between diet types. The levels of cis -aconitate, succinate, and histidine in wild-type mice were significantly higher than those in the AHNAK −/− mice in the chow diet group, but these were not significantly different in the HFD group. Levels of putrescine and threonine in AHNAK −/− mice were significantly higher in the HFD group than in the chow diet group. The increase in these metabolites led to a significant difference between the wild-type and AHNAK −/− mice in the HFD group. In addition, the levels of PAG and methionine in the different gene-type mice showed a reverse trend between the two diet types, that is, those metabolites in wild-type mice showed higher levels in the chow group, whereas those in the AHNAK −/− mice showed higher levels in the HFD diet group. 4 Discussion To examine the mechanism underlying the resistance to diet-induced obesity in AHNAK knock-out mice, we used 1 H NMR-based metabolomics to identify the diet-dependent changes in urinary metabolic profiles between wild-type and AHNAK −/− mice. In general, a large energy expenditure following activation of the tricarboxylic acid cycle (TCA) causes weight loss, so the TCA cycle intermediates appear to change even with minimal weight loss [16] and many studies have reported TCA cycle intermediates associated with a HFD and obesity in a rodent model [17–19] . In our study, the TCA cycle intermediates, such as succinate and cis -aconitate in the chow diet group, were higher in WT mice than in AHNAK −/− mice. However, the levels of those metabolites in the HFD-fed mice showed no significant difference between wild-type and AHNAK −/− mice. These data revealed that the resistance to high-fat diet-induced obesity in AHNAK −/− mice was not associated with an enhancement of the TCA cycle. Moreover, amino acids related to fatty acid metabolism were elevated in the urine of HFD-fed AHNAK −/− mice. First, threonine and histidine increased in the HFD-fed AHNAK −/− mice. Previous studies have demonstrated that these two amino acids are related to fat metabolism [20–23] . Yoshida and Harper [20] reported that fat synthesis is stimulated in rats fed a low-protein diet that is primarily deficient in threonine because the amounts of C-14 from injected acetate-l-C-14 incorporated into body and liver fat were significantly greater in rats fed the threonine-deficient diet than in those fed the threonine-supplemented diet. Furthermore, note that the parasite Trypanosoma brucei catabolizes l -threonine via a reaction involving threonine dehydrogenase and 2-amino-3-oxobutyrate CoA-ligase, converting threonine into glycine and acetyl-CoA and preferentially utilizing the acetyl-CoA produced for lipid synthesis, even in the presence of exogenous acetate [21] . Therefore, higher level of threonine in HFD-fed AHNAK −/− mice indicates that fat synthesis from threonine was suppressed in the AHNAK −/− . In addition, Kasaoka et al. [22] reported that dietary histidine suppresses food intake and fat accumulation in rats. In addition, Yoshimatsu et al. [23] reported that intraperitoneal administration of histidine accelerates lipolysis in white adipose tissue. In our study, the level of histidine was significantly higher in wild-type mice compared to AHNAK −/− mice in the chow diet group. However, the levels of histidine were significantly lower in the wild-type mice compared with HFD-fed wild-type mice, whereas that in HFD-fed AHNAK −/− mice were not significantly different compared to that in chow diet-fed the AHNAK −/− mice. Therefore, maintaining the levels of histidine between different diet types in AHNAK −/− mice may keep stimulating the lipolysis. Second, methionine and putrescine were elevated in the urine of the HFD-fed AHNAK −/− mice. Methionine, an essential amino acid belonging to the group of lipotropics [24] , helps eliminate fatty substances and prevent excessive fat from accumulating in the liver [25] . It is related to polyamine metabolism, producing a spermidine and spermine, which are polyamines. Putrescine, synthesized from ornithine by ornithine decarboxylase (ODC), is also a polyamine that is essential for the growth and normal function of cells. Therefore, higher levels of this metabolite indicate abundant polyamine flux. Recently, Jell et al. [26] showed the existence of a functional linkage between polyamine and fatty acid metabolism by SSAT (spermidine/spermine N 1-acetyltransferase). They also reported that elevated ODC and SSAT activities cause a diversion of acetyl-CoA from fatty acid synthesis to acetylation of polyamines, which are then excreted in the urine. That is, fatty acid oxidation was increased by ODC and SSAT activities which induced increasing polyamine flux, instead to fatty acid synthesis from acetyl-CoA, and it causes weight loss and a resistance to HFD-induced obesity. In addition, obesity itself is reportedly related to the gut bacteria [27] , and changes in metabolites that were affected by microflora, such as PAG and MA, showed different trends between wild-type and AHNAK −/− mice. Therefore, a major involvement of microflora occurs in the metabolic alteration of AHNAK −/− mice. However, further research is needed to clarify the relationship between those two metabolites and obesity. This study illustrates that AHNAK −/− mice have a resistance to diet-induced obesity and that 1 H NMR-based metabolomics can be used to distinguish the metabolic differences between wild-type and AHNAK −/− mice. In addition, the present study demonstrates that resistance to diet-induced obesity in AHNAK −/− mice is related to the perturbation of the amino acids that have an effect on fat metabolism and not from an enhancement of the TCA cycle. Therefore, global and targeted metabolic profiling via 1 H NMR-based metabolomics provides insight into understanding the biological pathways for the resistance to diet-induced obesity in AHNAK −/− mice. Acknowledgments This research was supported grants from National Research Foundation of Korea funded by the Korean Ministry of Education, Science and Technology ( 2009-008146 and PGA048) and Korea Basic Science Institute to Hwang GS; grant by Basic Science Research Program (No. R13-2008-028-01000-0 ) of the National Research Foundation of Korea funded by the Ministry of Education Science and Technology of Korea to Lee HS,; Research of New Drug Target Discovery by the Ministry of Education, Science and Technology of Korea to Seong J.K. (No. 20100020468 ) and Bae Y.S. (No. 20090093987 ). References [1] T. Hashimoto S. Gamou N. Shimizu Y. Kitajima T. Nishikawa Regulation of translocation of the desmoyokin/AHNAK protein to the plasma membrane in keratinocytes by protein kinase C Exp. Cell Res. 217 1995 258 266 [2] H. Haase T. Podzuweit G. Lutsch A. Hohaus S. Kostka C. Lindschau M. Kott R. Kraft I. Morano Signaling from beta-adrenoceptor to L-type calcium channel: identification of a novel cardiac protein kinase A target possessing similarities to AHNAK FASEB J. 13 1999 2161 2172 [3] F. Sekiya Y.S. Bae D.Y. Jhon S.C. Hwang S.G. Rhee AHNAK, a protein that binds and activates phospholipase C-gamma1 in the presence of arachidonic acid J. Biol. Chem. 274 1999 13900 13907 [4] I.Y. Kim J.H. Shin J.K. Seong Mouse phenomics, toolbox for functional annotation of human genome BMB Rep. 43 2010 79 90 [5] A. Komuro Y. Masuda K. Kobayashi R. Babbitt M. Gunel R.A. Flavell V.T. Marchesi The AHNAKs are a class of giant propeller-like proteins that associate with calcium channel proteins of cardiomyocytes and other cells Proc. Natl. Acad. Sci. USA 101 2004 4053 4058 [6] M. Kouno G. Kondoh K. Horie N. Komazawa N. Ishii Y. Takahashi J. Takeda T. Hashimoto Ahnak/Desmoyokin is dispensable for proliferation, differentiation, and maintenance of integrity in mouse epidermis J. Invest. Dermatol. 123 2004 700 707 [7] R.L. Byyny M. LoVerde S. Lloyd W. Mitchell B. Draznin Cytosolic calcium and insulin resistance in elderly patients with essential hypertension Am. J. Hypertens. 5 1992 459 464 [8] D. Matza A. Badou K.S. Kobayashi K. Goldsmith-Pestana Y. Masuda A. Komuro A scaffold protein, AHNAK1, is required for calcium signaling during T cell activation Immunity 28 2008 64 74 [9] B. Draznin K. Sussman M. Kao D. Lewis N. Sherman The existence of an optimal range of cytosolic free calcium for insulin-stimulated glucose transport in rat adipocytes J. Biol. Chem. 262 1987 14385 14388 [10] B. Draznin K.E. Sussman R.H. Eckel M. Kao T. Yost N.A. Sherman Possible role of cytosolic free calcium concentrations in mediating insulin resistance of obesity and hyperinsulinemia J. Clin. Invest. 82 1988 1848 1852 [11] J.K. Nicholson J.C. Lindon E. Holmes Metabonomics: understanding the metabolic responses of living systems to pathophysiological stimuli via multivariate statistical analysis of biological NMR spectroscopic data Xenobiotica 29 1999 1181 1189 [12] L.M. Raamsdonk B. Teusink D. Broadhurst N. Zhang A. Hayes M.C. Walsh J.A. Berden K.M. Brindle D.B. Kell J.J. Rowland H.V. Westerhoff K. van Dam S.G. Oliver A functional genomics strategy that uses metabolome data to reveal the phenotype of silent mutations Nat. Biotechnol. 19 2001 45 50 [13] I.H. Lee H.J. Lim S. Yoon J.K. Seong D.S. Bae S.G. Rhee Y.S. Bae Ahnak protein activates protein kinase C (PKC) through dissociation of the PKC-protein phosphatase 2A complex J. Biol. Chem. 283 2008 6312 6320 [14] T. Skov F. van den Berg G. Tomasi R. Bro Automated alignment of chromatographic data J. Chemometr. 20 2006 484 497 [15] L. Eriksson, E. Johansson, N. Kettaneh-Wold, S. Wold, Multi-and megavariate data analysis, Part I, Umetrics Umea, Sweden, 2006. [16] S.C. Connor W. Wu B.C. Sweatman J. Manini J.N. Haselden D.J. Crowther C.J. Waterfield Effects of feeding and body weight loss on the 1 H NMR-based urine metabolic profiles of male Wistar Han rats: implications for biomarker discovery Biomarkers 9 2004 156 179 [17] J. Shearer G. Duggan A. Weljie D.S. Hittel D.H. Wasserman H.J. Vogel Metabolomic profiling of dietary-induced insulin resistance in the high fat-fed C57BL/6J mouse Diab. Obes. Metab. 10 2008 950 958 [18] K. Sheikh G. Camejo B. Lanne T. Halvarsson M.R. Landergren N.D. Oakes Beyond lipids, pharmacological PPARalpha activation has important effects on amino acid metabolism as studied in the rat Am. J. Physiol. Endocrinol. Metab. 292 2007 E1157 E1165 [19] R.E. Williams E.M. Lenz J.A. Evans I.D. Wilson J.H. Granger R.S. Plumb C.L. Stumpf A combined (1)H NMR and HPLC-MS-based metabonomic study of urine from obese (fa/fa) Zucker and normal Wistar-derived rats J. Pharm. Biomed. Anal. 38 2005 465 471 [20] A. Yoshida A.E. Harper Effect of threonine and choline deficiencies on the metabolism of C14-labeled acetate and palmitate in the intact rat J. Biol. Chem. 235 1960 2586 2589 [21] R.A. Klein D.J. Linstead Threonine as a perferred source of 2-carbon units for lipid synthesis in Trypanosoma brucei Biochem. Soc. Trans. 4 1976 48 50 [22] S. Kasaoka N. Tsuboyama-Kasaoka Y. Kawahara S. Inoue M. Tsuji O. Ezaki H. Kato T. Tsuchiya H. Okuda S. Nakajima Histidine supplementation suppresses food intake and fat accumulation in rats Nutrition 20 2004 991 996 [23] H. Yoshimatsu K. Tsuda A. Niijima M. Tatsukawa S. Chiba T. Sakata Histidine induces lipolysis through sympathetic nerve in white adipose tissue Eur. J. Clin. Invest. 32 2002 236 241 [24] C.H. Best J.H. Ridout The lipotropic action of methionine J. Physiol. 97 1940 489 494 [25] G.S. Lee J.S. Yan R.K. Ng S. Kakar J.J. Maher Polyunsaturated fat in the methionine-choline-deficient diet influences hepatic inflammation but not hepatocellular injury J. Lipid Res. 48 2007 1885 1896 [26] J. Jell S. Merali M.L. Hensen R. Mazurchuk J.A. Spernyak P. Diegelman N.D. Kisiel C. Barrero K.K. Deeb L. Alhonen M.S. Patel C.W. Porter Genetically altered expression of spermidine/spermine N1-acetyltransferase affects fat metabolism in mice via acetyl-CoA J. Biol. Chem. 282 2007 8404 8413 [27] P.J. Turnbaugh R.E. Ley M.A. Mahowald V. Magrini E.R. Mardis J.I. Gordon An obesity-associated gut microbiome with increased capacity for energy harvest Nature 444 2006 1027 1031