Pyruvate
carboxylase
in genetic obesity
CHRISTOPHER J. LYNCH, KENNETH M. MCCALL, MELVIN L. BILLINGSLEY, LIZABETH M. BOHLEN, STANLEY P. HRENIUK, LOUIS F. MARTIN, LEE A. WITTERS, AND SUSAN J. VANNUCCI Departments of Cellular and Molecular Physiology, Pharmacology, and Surgery, College of Medicine, The Pennsylvania State University, Hershey, Pennsylvania 17033; and Endocrine-Metabolism Division, Departments of Medicine and Biochemistry, Dartmouth Medical School, Hanover, New Hampshire 03756 Lynch, Christopher J., Kenneth M. McCall, Melvin L. Billingsley, Lizabeth M. Bohlen, Stanley P. Hreniuk, Louis F. Martin, Lee A. Witters, and Susan J. Vannucci. Pyruvate carboxylase in genetic obesity. Am. J. Physiol. 262 (Endocrinol. Metab. 25): E608E618, 1992.-Immunoblotting and protein microsequencingwere usedto identify severaladipocyte proteins expressedin an obesity-related fashion in the Zucker rat. One of these was a 116-kDa particulate protein (~116). The ~116 levelsin adipocytes from 5 to 7-wk-old obese Zucker rats were two- to fivefold higher on a per milligram of protein basisthan levelsin lean animalsand decreasedafter the induction of streptozotocin-induced diabetesmellitus. This suggests the change may be related to the actions of insulin. Hepatic levelsof ~116 did not change.The ~116 waspurified to homogeneity from obese Zucker rat adipocytes, and polyclonal antisera were prepared against the purified protein in rabbits. Microanalysis of electroblotted pII6 proteolytic fragmentssuggestedthat ~116waspyruvate carboxylase(PC). Other evidence that ~116 was PC included the following: 1) ~116 contained biotin, 2) ~116 in particulate subcellularfractions was soluble after freeze-lysis, 3) antibodies to ~116 reacted with purified hepatic PC, 4) ~116 and purified hepatic PC had identical p1 and relative molecular weight values, and 5) similar changes weredetectedin adipocyte ~116 and PC enzyme activity during obesity and after the induction of streptozotocin-induced diabetesmellitus. IncreasedadiposetissuePC probably contributes to the increasedlipogenic capacity of young obeseZucker rat adipocytes. fatty acid synthetase; adenosine5’-triphosphate-citrate lyase; carbonic anhydrase III; coenzyme A carboxylase; Zucker rat; diabetesmellitus; lipogenesis;avidin; insulin; microsequencing
betweenhumanobesity, presumed to have some genetic component, and the characteristics of Zucker fa/fa obese rats. It is not surprising, therefore, that these animals have been extensively studied regarding the biochemical, physiological, and behavioral manifestations of genetic obesity (for reviews see Refs. 3, 9, 16, 22). Important information has been gained from these studies on the biochemical changes in the adipocyte and other tissues during the development and maintenance of the obese state. Despite this, a complete picture of the disease etiology is not yet available, especially regarding the identification of the fa gene or genes. To address this problem, our laboratory is mapping and identifying specific proteins differentially expressed during obesity in adipose and
THEREAREMANYSIMILARITIES
other tissues of genetically obese Zucker rats.
In this communication, we confirm the increased expression of the known lipogenic enzymes, fatty acid synthetase, ATP-citrate lyase, and acetyl-CoA carboxylase previously reported in this model. We also report on the purification and identification of a protein whose expression increases two- to fivefold in adipocytes from E608
0193-1849/92
young obese Zucker rats and decreases during strepto-
zotocin-induced diabetes mellitus in these animals. The data suggest that this protein is pyruvate carboxylase (PC). Given the critical role of PC in maintaining tricarboxylic acid cycle intermediates (14, 18, 27, 32-33), which are presumably being actively drawn off in obesity for glycerogenesis and lipogenesis, we propose that PC contributes to the hypertrophy of adipose tissue during the development of obesity. EXPERIMENTAL PROCEDURES Isolation of adipocytes, subcellular fractions, and liver homogenates. Intact adipocyteswerepreparedfrom the parame-
trial fat pads of anesthetized and fasted lean and obesefemale Zucker rats as previously describedin detail (41). Isolated fat cellswere processedeither for gel electrophoresisor subcellular fractionation. For gel electrophoresis,isolated fat cells were washedin Krebs-Ringer bicarbonate buffer (pH 7.4) containing 200 nM adenosinebut without the usual4% bovine serumalbumin fraction V. These concentrated cellswere frozen at -85°C as 1503.Jaliquots and were assayedfor protein content using Pierce CoomassieBlue reagent. For subcellular fractionation, cells were resuspendedin 10 vol of ice-cold homogenization buffer (IO mM Na2P04, 5 mM EDTA, pH 7.4, including 5 pg/ml leupeptin, 5 pg/ml soybean trypsin inhibitor, and 5 pg/ml aprotinin). The cells were homogenized at 4°C in a PotterElvehjem homogenizerand centrifuged at 10,000g to isolate a low-speedparticulate subcellular fraction containing plasma membranevesicles,mitochondria, and nuclei. These fractions were washed during a second 10,000g centrifugation, resuspendedin homogenizationbuffer, assayedfor protein asabove, and stored at -85°C. The 10,000-gsupernatant wascentrifuged at 100,000g for 1 h to isolatethe high-speedparticulate subcellular fraction, which wasthen washedin homogenizationbuffer for 1 h at 100,000g. The 100,000-gpelletswere similarly assayed for protein content and stored at -85°C in homogenization buffer. The 100,000-gsupernatant, termed the cytosolic fraction, wasconcentrated without significantly increasingsalt levelsby vacuum dialysis against 4 liters of homogenizationbuffer in a Bio-Molecular Dynamics (Beaverton, OR) Micro-ProDiCon device. Aliquots of the concentrated cytosolic fraction were assayedfor protein, and the remainder was frozen at -85°C. Livers from these animalswere briefly perfusedand homogenized in ice-cold homogenizationbuffer using a Tekmar tissue homogenizerand stored in aliquots at -85°C. Gel electrophoresis. One-dimensionalsodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and Coomassie Blue staining were performed as previously described (21) using a Hoefer SE600 apparatus with 4°C cooling. Twodimensional isoelectric focusing (IEF)-SDS-PAGE and silver staining weredoneaccordingto Hochstrasseret al. (19). Briefly, adipocyte or liver proteins were diluted to 1.5 mg/ml in 10% SDS and 23 mg/ml dithiothreitol. These mixtures were solubilized using a Tekmar Vibracell sonicatorfitted with a 2-mm microtip probe and mixed in a 1:4 ratio with a solution contain-
$2.00 Copyright 0 1992 the American Physiological
Society
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
PROTEIN PROTEI N EXPRESSION
1
2
IN GENETIC
34
56
116~
Fig. 1. Adipocyte proteins from lean and obese Zucker rats. Adipocytes (70 kg protein/lane) and adipocyte low-speed particulate subcellular fractions (50 fig protein/lane) prepared from parametrial fat pads of 5- to 7-wk-old heterozygous lean and homozvsous obese Zucker rats. as described in text, were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) and Coomassie Blue staining. Lane 1, adipocytes from lean rats; lane 2, adipocytes from obese rats; lanes3 and 4, lowspeed particulate fractions from lean rats; lanes 5 and 6, low-speed particulate fractions from obese rat adipocytes. Arrow indicates 116-kDa protein band (~116) M,, relative molecular weight.
67-
cy) 10 l-
E609
OBESITY
“-
43-
X
nrlinnnzrtn
dye and waselectrophoresedat per gel until the tracking dye migrated to the bottom of the Zucker ruts as compared with lean lit(term&es resolving gel. Protein Concentration in Obese flm3ntification Purification of ~116, antibody production, and Western Designation Adipocytes Relative to Lean blotting. Polyclonal antibodies to ~116 were preparedin rabbits ~265 (C) elevated (19-fold higher) Acetyl-CoA carboxylase* as follows. The ~116 was purified from 10,000-gparticulate ~260 (C) elevated (19-fold higher) Fatty acid synthetase* fractions of obeseZucker rat adipocytes using preparative (100 ~116 (P) elevated (2- to 5-fold higher) Pyruvate carboxylase rugprotein/tube) two-dimensional IEF-SDS-PAGE. The ~116 ~112 (C) elevated (19-fold higher) ATP-citrate lyase* ~106 (P) decreased (80-90% less) Unknown gelspot centerswereremovedfrom CoomassieBlue-stainedgels decreased (80-90% less) Unknown in destain usinga 3-mm coring tool. Twelve of thesespotswere decreased (40-60% less) Carbonic anhydrase III? collectedper iniection. Gel snotswere eouilibrated for 10 min in Seven major proteins (p) were initially observed to be differentially 10ml of water,placed in microfugetube;, frozen in liquid nitroexpressed in a series of one-dimensional gel electrophoresis experiments gen, and lyophilized in a Jouvan RC 10.10 centrifugal freeze such as that shown in Fig. 1. P, particulate fraction; C, cytosolic dryer. Dried gel spots were powdered with a microfuge tube fraction. See text for explanation of pyruvate carboxylase. * Identified homogenizing pestle and resuspendedin 1.5 ml of sterile isoin immunoblot studies (not shown) using previously described antisera (6, 12). Changes in acetyl-CoA carboxylase mass were also confirmed tonic phosphate-bufferedsaline. This gel mixture wasdrawn up by enzyme-linked immunosorbent assay, as previously described (20). into a 5-ml luer-lock glasssyringe containing 1.5 ml of either tconfirmed by microsequencing isoform-specific antisera and enzy- Freund’s completeadjuvant (Sigma), Freund’s incompleteadjumetic activity measurements (C. J. Lynch, K. M. McCall, R. L. vant, or 8 mg/ml of aluminum hydroxide adjuvant. These mixHoretsky, S. J. Vanucci, N. Carter, and S. J. Dodgson, unpublished tures wereemulsifiedusing a stainlesssteelmale-maleleur-lock observations). microemulsifying tube (ZO-guageX 7/8 in.; Thomas Scientific) ing 4% 3-[(3-cholamidopropyl)dimethylammonio]-l-propane- connected to another 5-ml leur-lock tube. The mixture was sulfonate detergent, 54% urea, 10 mg/ml dithiothreitol, and 4% injected subcutaneously and intradermally at multiple sites ampholines[a 4:l mixture of BDH Chemicals(Poole, UK) pH along the shavedback of a male New Zealandrabbit (~116with 3-10 and pH 5-7 40% ampholine solutions]. This mixture was Freund’s adjuvant) or intraperitoneally (~116 with aluminum sonicated, and 50 ~1 were applied to 120 x 1.5 mm IEF poly- hydroxide adjuvant). The injection, adjuvant, and bleeding acrylamide (4% acrylamide) tube gelsusing diacrylylpiperazine schedulewasthat of Green et al. (15). For humanepurposesand (BioRad) asa cross-linker instead of N,N’-methylene-bis-acry- as a bleedingaid, rabbits were anesthetizedusing a Droperidollamide (19). Sampleswere electrofocusedin a Hoefer tube gel Fentanyl mixture (Innovar-Vet) during every procedure.Immuapparatusfor 7,500 V hours at room temperature according to noblotting was performed according to Lynch et al. (26) using the following program: 200V for 2 h, 400 V for 15 h, and 800 V i”“I-labeled goat anti-rabbit immunoglobulin G F’Ab, fragfor the remainder.Tube aelswereeentlv extruded usinadistilled ments and autoradiography as a meansfor identifving antibodv water, blotted briefly without touihingthe gel,and gently rolled binding proteins or, as describedby Bianchi et al: (6r, for fatty between two glass plates filled with SDS-PAGE resolving acid synthetase and acetyl-CoA carboxylase. For detection of buffer, the 4% stacking and 7.5% resolving SDS-PAGE. The biotin-containing proteins on Western blots, we used avidintube gel in contact with the 4% gel was covered with 1 ml of linked alkaline phosphatase,as describedby Billingsley et al. Laemmli sample buffer (21), containing bromophenol blue (7). Tahlo 1_cIUIL.
1I.
Phnnnnc “rYLMv~6co
;n “I”
mninr rr”uJ”r
ULufJ”LJY~,
vrot&s
in
,-&se
tracking
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
E610
PROTEIN
EXPRESSION
IN GENETIC
IEF-PAGE
OBESITY
IEF-PAGE
OBESE
LEAN ADIPOCYTES
-116 -97.4 -67
:
a a I
-43
iz co
-29
OBESE
LEAN LIVER
Fig. 2. Proteins from adipocyte lysates and liver homogenates. Adipocyte lysates (17 wg protein/tube gel) and liver homogenates (17 pg protein/tube) subjected to two-dimensional isoelectric focusing (IEF)-SDS-PAGE and silver staining. Top left: adipocyte lysate from lean rats. Top right: adipocyte lysate from obese rats. Bottom Left: liver homogenate from lean rats. Bottom right: liver homogenate from obese rats. Arrow indicates ~116 (p1 = 7.0, size = 116,000 kDa).
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
PROTEIN
EXPRESSION
IEF-PAGE
IN GENETIC
OBESITY
E611
100KCl, 50 tris(hydroxymethyl)aminomethane.HCl (pH 8.1 at 37”C), 15 [14C]KHC03, 2.4 ATP, 0.2 acetyl-CoA, 5 MgC&, 0.1 ethylene glycol-bis(&aminoethyl ether)-N,N,N’,N’-tetraacetic acid, 1 NADH, and 65 U of malate dehydrogenase(Boehringer Mannheim; “special quality”) in the presenceor absenceof 10 mM sodiumpyruvate in a total volume of 1 ml. Reactionswere stopped with the injection of 1 ml of 2 N sodium acetate (pH 3.4). This mixture wasshakenat 37°C for 1 h. After that period, the center well containing hyamine hydrochloride asa COstrap was removed, and the radioactive content of the reaction mixtures was determined using a Beckman LS-3801 liquid scintillation counter. PC waspurified from rat liver using the method of Ahmad et al. (1). The specificactivity of the purified enzyme was31 U/mg protein and was stored (-84°C) at a concentration of -1.0 mg protein/ml. RESULTS
Fig. 3. Low-speed particulate adipocyte proteins from obese rat. Silver stain of two-dimensional IEF-SDS-PAGE of low-speed particulate subcellular fraction from obese Zucker rat adipocytes.. Such gels subjected to reversible Coomassie Blue staining were used as source of ~116 (arrow). Internal amino acid sequence analysis of ~116. Limited proteolysis of purified p 116 by the Clevelandet al. (11) method was obtained when’ Staphylococcus aureus V8 protease (l-2 ng protein/gel lane) and ~116 (lo-20 Kgprotein/lane in gel circles isolatedasdescribedabove) were allowedto electrophoreseinto a 4% acrylamide stacking gel poured over a 12% acrylamide SDS-PAGE resolving gel and then allowedto incubate at 30°C for 30 min without current. After the digestion period, electrophoresis resumedat 25 mA constant current, and the buffer chamber was cooled back to 4°C. When the tracking dye reached the bottom of the 12% acrylamide SDS-PAGE, peptides were transferred to polyvinylidene difluoride (PVDF) membranesat 500 mA constant current for 90 min in transfer buffer containing 10 mM 3-(cyclohexylamino)-l-propanesulfonic acid (pH 11) and 10% methanol, asdescribedby Matsudaira(29). The blotting apparatus (BioRad) contained a cooling core through which 4°C coolant was circulated. Visualization of proteins and peptideswasby CoomassieBlue staining (29). Bands containing peptides were cut with a clean razor blade, and sequenceanalysis was performed using an Applied Biosystemsmodel470-A sequenatorequippedwith on-line phenylthioisocyanate analysis, as describedby Matsudaira (29). PC activity and purification. PC activity was assayedunder conditions needed for maximal activity, as described by McClure et al. (30-31) using a radioisotopic method for the detection of pyruvate-dependent [ 14C]bicarbonatefixation as follows: 25 ~1of samplewere incubated at 37°C with (in mM)
In initial one-dimensional gel electrophoresis experiments, such as those depicted in Fig. 1, we reproducibly observed differential staining of seven major proteins from adipocytes of obese Zucker rats compared with lean littermates on a per milligram protein basis (Table 1). To identify these proteins, subcellular fractionation studies using isopycnic centrifugation (Fig. 1 and data not shown) and metabolic radiolabeling experiments employing inorganic [32P]PO: were initially performed (C. J. Lynch, K. M. McCall, R. L. Horetsky, S. J. Vanucci, N. Carter, and S. J. Dodgson, unpublished observations). This data together with the pI and relative molecular weight (M,) of the proteins, as determined by us, was used to computer search adipocyte, Zucker rat, and 3T3 cell line literature. From these efforts the identities of three of the cytosolic (i.e., 100,000-g supernatant) proteins with relative molecular sizes of -265, 260, and 112 kDa, were suggested, i.e., acetyl-CoA carboxylase (p265), fatty acid synthetase (p260), and ATP-citrate lyase (~112). This was confirmed by quantitative immunoblotting studies using antisera that specifically recognized acetyl-CoA carboxylase and antisera that recognized both fatty acid synthetase and ATP-citrate lyase (Table 1). The identification of ~112 in our gels was further substantiated in other experiments (see below) with ATP-citrate lyasespecific antibodies kindly supplied by Dr. Charles Rubin (Albert Einstein College of Medicine, New York, NY). The identity of the other four proteins could not be so determined. These included three particulate proteins of 116, 106, and 88 kDa and a 28-kDa cytosolic protein, termed ~116, ~106, ~88, and ~28, respectively. Three of the proteins, ~106, ~88, and ~28, displayed decreased staining when SDS-PAGE of adipocyte proteins from obese Zucker rats was compared with lean littermates (Fig. 1 and Table 1). The ~28, an abundant protein in lean rodent adipose tissue, has been recently purified and identified by microsequencing as well as two-dimensional immunoblotting as carbonic anhydrase III (C. J. Lynch, K. M. McCall, R. L. Horetsky, S. J. Vanucci, N. Carter, and S. J. Dodgson, unpublished observations). Purification and identification of p106 and p88 is now proceeding in our laboratory. Staining of the other, a li6-kDa particulate protein (~116, the primary subject of the present study), was increased two- to fivefold on a per milligram protein basis in adipocytes from obese Zucker rats (Fig.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
rats were subjected to subcellular fractionation (50 pg protein/lane) and SDS-PAGE. Proteins were transferred from gels to Nitroplus transfer membrane (Micron Separations, Westboro, MA) and subjected to immunoblot analysis using lz51-labeled anti-rabbit gamma-globulin F’Ab2 fragments followed by autoradiography of resulting blots. Relevant molecular weight region of immunoblots from duplicate gels. Lane I, whole cell lysate; lane 2, low-speed particulate subcellular fraction; lane 3, high-speed particulate subcellular fraction; and lane 4, concentrated high-speed supernatant. A: antisera was raised against purified ~116 (1:200 dilution). B: antisera was raised against purified ATP-citrate lyase (1:2000 dilution) and was a generous gift of Dr. Charles Rubin, Albert Einstein College of Medicine, New York, NY.
Table 2. Amino acid alignment of internal sequences from ~116 with yeast PC Item
Species
Homology with p116, %
Aligned Sequence
24K peptide
Rat
LFQLXPAQNRAQKLL
23K peptide
Rat
LFQLRP AQNRAQKLLHYLG
PC
Yeast
20K peptide
Rat
PC
Yeast
18K peptide
Rat
PC
Yeast
lOT
70
lOT
LFQMV: ZSQNRAQKLLHYLA 480T
AAISYXl SXVADPSR
lOT
25
ALKDVI (GQIGAPMP 301 KIGINGXAIQXRVQAEXPA
63
KITTRGFAIQCRITTEDPA
lOT
360T
* Purified ~116 was subjected to limited proteolysis using Stuphylococcus aureus V8 protease. Several peptides on polyvinyl difluoride Western blots resulting from this digestion were subjected to automated gas-phase sequencing as described in text. They are compared with databank sequences submitted by Lim et al. (25) and deduced from a cloned fragment of genomic DNA containing pyruvate carboxylase yeast (PC) gene. Percent homology refers to % exact matches excluding Xs (unrecognizable or ambiguous amino acid in 2 separate sequencing attempts). Decapeptide marks appear below sequences.
1). In two-dimensional analytic gels (17 pg protein/tube), a single 116-kDa protein could be observed, which had a p1 of 7.0 and displayed increased staining in gels from obese relative to lean animals (Fig. 2, top). A protein with identical mobility in two-dimensional gels was also seen in livers from Zucker rats (Fig. 2, bottom). The abundance of the 116-kDa liver protein was not influenced by obesity, however. The ~116 purification
and immunoblotting studies. We
purified ~116 by preparative two-dimensional IEF-SDSPAGE for polyclonal antisera production and protein sequencing. Because ~116 seemed to be most abundant in low-speed particulate fractions from obese Zucker rats, as determined from one-dimensional SDS-PAGE and Coo-
massie Blue staining (data not shown), this preparation was used as starting material for purification by preparative two-dimensional IEF-SDS-PAGE. Figure 3 shows a representative preparative two-dimensional IEF-SDSPAGE of adipocyte particulate proteins from obese Zucker rats. The most abundant protein in such gels was P 116 (i.e., p1 = 7.0, M, = 116,0&O), as detected by CcIOmassie Blue (data not shown) or silver staining (Fig. 3). The centers of such spots from Coomassie Blue-stained gels, which should theoretically contain only pure ~116 protein, were used to prepare polyclonal antibodies in rabbits and for microsequencing studies. The antisera were used in two-dimensional IEF-SDS-PAGE immunoblotting experiments to show that the protein identified in liver was ~116. Figure 4 shows that the immunoreactivity of our antisera is against ~116, since the 116-kDa immunoreactivity is most abundant in adipocyte particulate fractions and is increased fivefold in Zucker rat obesity, as opposed to ~112 (ATP-citrate lyase) immunoreactivity, which is localized in the cytosolic fractions and is increased -19-fold in fractions from obese relative to lean animals.
Studies to determine the identity of ~116. Coomassie Blue-stained PVDF blots were obtained from ~116 digests using Staphylococcus aureus V8 protease for limited digestion and SDS-PAGE containing 12% acrylamide for peptide separation. Such digestions yielded four peptides that did not comigrate with peptides found in lanes with protease alone (data not shown). The peptides had relative molecular sizes of 24, 23, 20, and 18 kDa. These were subjected to gas-phase microsequencing in two separate experiments using different batches of protease and ~116 (Table 2). The 24- and 23-kDa peptides appeared identical at their NHs-termini, with the exception that, in two separate experiments, ambiguous sequence information was obtained from the 24-kDa peptide after the fifteenth sequencing cycle. The alignment routine of the DNA-STAR computer program was used to search the National Biomedical Research Foundation Protein Identification Resource for matching amino acid sequences. The best overall homology was with PC from
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
PROTEIN
EXPRESSION
IN GENETIC
IEF-PAGE
12
W (3 a a A CI co
v-”\ h--i : ‘j/ 1 I _i
.,‘
/*
/ : -_ >r
E613
OBESITY
/ - I/’_ .‘;Ci’2/(
\-i3-%a - I I i
34
?/iYi ’ \ 1tI \ / 51 : iI i / \** 1; L/ // jj is i ;? 3: /’ : \i I x : i i \ /i n * * 3* , , -\ \\ _ tt 3 , ’ I ,I1’r*I /
Fig. 5. Avidin binding proteins in adipocytes and liver. Proteins on Western blots binding avidin conjugated to biotinylated alkaline phosphatase were visualized using nitro blue tetrazolium chloride-5-bromo-4-chloro-3-indoyl phosphatep-toluidine salt chromagen system (7). Left: low-speed particulate subcellular fractions from obese Zucker rat adipocytes were separated by two-dimensional IEF-SDS-PAGE. Arrow indicates position of ~116 (compare with stained gel in Fig. 3). Right: SDS-PAGE of adipocyte lysates from lean (lane I) and obese (lane 2) Zucker rats or liver homogenates from lean (lane 3) and obese (lane 4) Zucker rats.
yeast (25), having between 25 and 70% exact matches with internal peptides. To compare our sequences with those of full-length mammalian PC sequences, several other data banks (i.e. Swiss Protein data bank release 15, European Molecular Biology Laboratories data bank release 24, and GenBank release 65) were searched using either the term PC or EC 6.4.1.1. These searches did not reveal published full-length mammalian cDNA encoding PC which could be translated. The homotetrameric subunits of yeast PC each have a calculated molecular mass of 116 kDa and a calculated p1 of 7.0, i.e., like ~116 (25). However, we were not entirely convinced that ~116 was PC in the absence of a translatable full-length mammalian cDNA. One reason for contern was that earlier studies on mammalian PC reported a relative molecular mass of - 130 kDa (e.g., Refs. 13, 14, 31), somewhat higher than ~116. Several added experiments were performed to test the hypothesis that ~116 was PC. PC contains a biotin moiety that transfers bicarbonate to form oxaloacetate. The protein avidin tightly binds to biotin residues (7), and avidin conjugated to a chromogen-generating system was used to look for highaffinity avidin binding by ~116, as described in EXPERI-
PROCEDURES. Figure 5 (left> shows that ~116 isolated in two-dimensional IEF-SDS-PAGE and immobilized on nitrocellulose membranes does bind an avidin conjugate. This 116-kDa avidin binding activity was less abundant in liver homogenates compared with adipocytes and was increased in adipocytes from obese relative to lean Zucker rats (Fig. 5, right). This suggests that activity represented binding to biotin attached to ~116 and not binding to some other biotin-containing 116-kDa protein, PC is a soluble protein but is found almost exclusively inside mitochondria and therefore would not be expected to be found in cytosolic fractions (3). Manipulations that break mitochondria release a soluble form of PC. The ~116 was found in low-speed particulate subcellular fractions containing mitochondria but not cytosol (Fig. 4). However, it was unclear whether ~116 was a soluble protein, an integral membrane, or cytoskeletal protein, Figure 6 shows that, when freeze-lysed 10,000-g pellets were sonicated in hypotonic buffer to turn mitochondria inside out, ~116 was released in a soluble form which would not sediment at 100,000 g. Figure 7 shows that ~116 antibodies react with PC purified from rat liver, and Fig. 8 shows that purified MENTAL
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
E614
PROTEIN
EXPRESSION
IN GENETIC
OBESITY
1
1
-pi /,
16
Fig. 6. Release of soluble ~116 from low-speed particulate subcellular fractions after freeze-lysing and sonication in hypotonic media. Freshly prepared low-speed particulate subcellular fractions from obese Zucker rats, diluted to 1 ml (1 mg protein/ml), were frozen in liquid Ns, thawed, and sonicated. This suspension (50 ~1) was reserved (lane 1) and remainder was centrifuged at 100,000 g for 1 h. Supernatant (50 ~1) was removed (lane 2), and both samples were subjected to SDS-PAGE followed by Coomassie Blue staining.
hepatic PC and ~116 comigrate in two-dimensional IEF-SDS-PAGE. Last, Fig. 9 shows that PC activity was increased in young obese Zucker rat adipocytes when these were compared with adipocytes from lean controls. The increase in activity was consistent with the increases in ~116 immunoreactivity, avidin binding, and staining, which ranged from two- to fivefold in our studies. All of the above findings support the hypothesis that ~116 is PC. Effect of streptozotocin-induced diabetes mellitus on PC. Caro et al. (10) have forwarded a model of human obesity in which hyperinsulinemia and insulin resistance, progressing with age more slowly in adipose tissue vis-a-vis other peripheral tissues, result in preferential shunting of substrates to adipose tissue resulting in adipocyte hypertrophy and adipose tissue hyperplasia. Consistent with Caro et al.3 model, young obese Zucker rats display hyperinsulinemia and increased adipose tissue responsiveness to insulin (9, 17). To determine the possible role of hyperinsulinemia in the obesity-related increases in PC that we observed, the effects of streptozotocin-induced diabetes mellitus on adipocyte PC concentrations and activities in the obese Zucker rat were studied (Figs. 9-10). Figure 9 shows that acute diabetes mellitus reversed the effects of obesity on PC activity in
Fig. I. Recognition of uepa~~ pyruva~ cdruoay~x antibodies. Purified hepatic PC was subjected to SDS-I m-r!, on IWO acrylamide gel. One-half of gel containing 2 pg of purified hepatic PC was Coomassie Blue stained (lane 1). Other half (lane 2) also containing 2 pg of purified hepatic PC was transferred to Nitroplus transfer membrane and subjected to immunoblot analysis using ~116 antibodies, as
describedin Fig’ 4’ adipocytes. Figure 10 shows that concentrations of PC (~116) decreased in acutely diabetic obese animals. Concentrations of ~265, ~260, and ~112 also decreased in diabetic obese rats while p28 increased. The ~106 and ~88, major proteins in adipocytes from lean animals (Fig. l), were not readily observed in Coomassie Blue-stained SDS-PAGE of either obese or diabetic obese rat adipocyte proteins (Fig. 10). These findings support the hypothesis that PC mass and activity increase in adipocytes from young obese Zucker rats and that such changes may be related to the actions of insulin. DISCUSSION Seven major adipocyte proteins and phosphoproteins were initially observed to be altered in obesity (Fig. 1 and
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
? co
IEF-PAGE
I
C
Fig. 8. Comigration of purified hepatic PC and adipocyte ~116 in twodimensional IEF-SDS-PAGE. Proteins from low-speed particulate subcellular fraction from obese Zucker rat adipocytes containing ~116, indicated by arrow (A), purified hepatic PC (B), or 1:l mixture of both (C) were subjected to two-dimensional IEF-SDS-PAGE and silver staining. STD, lane of standard molecular weight markers.
! _ “/
” i*buer*w*n-e - NUwrVI_
Table 1). From previous studies on adipocytes or 3T3 cell lines (e.g., 5, 8, 23, 28, 36-39), we deduced and subsequently confirmed the identity of three of these by immunoblotting to be acetyl-CoA carboxylase (p265), fatty acid synthetase (p260), and ATP-citrate lyase (~112) (Table 1, Fig. 4, and data not shown). The direction of the change in the expression of these proteins in the obese Zucker rats is in keeping with previous studies on their activity in this animal model (5, 23, 28, 37, 39). However, the magnitude of the change determined in this study was higher, i.e., 19-fold. One possible explanation for this difference is that the activity of some of these enzymes is regulated by phosphorylation (34) and other mechanisms (2, 24) that may be preserved during cell breakage and enzyme assay. Our findings are the first evidence that the actual mass of these enzymes is increased in the obese Zucker rat. The identification of the other four adipocyte proteins found altered in obese Zucker rats was not so determined. Purification and identification of two particulate proteins, namely ~106 and ~88, is now under way. A
third, ~28, has been identified by protein microsequencing, immunoblotting, and other experiments as carbonic anhydrase III (C. J. Lynch, K. M. McCall, R. L. Horetsky, S. J. Vanucci, N. Carter, and S. J. Dodgson, unpublished observations). The present communication describes the purification and identification of the fourth, ~116. Coomassie Blue staining of ~116, a ll6-kDa major adipocyte protein, was increased between two- and fivefold in SDS-PAGE of adipocyte proteins from obese compared with lean Zucker rats (Figs. 1-2). A liver protein that was identical to ~116 in terms of two-dimensional IEF-SDS-PAGE mobility did not change in response to obesity (Fig. 2). The ~116 protein was purified to homogeneity by preparative two-dimensional IEF-SDS-PAGE using high-density adipocyte membranes from obese Zucker rats as starting material (Fig. 3). Polyclonal antibodies were prepared against the purified protein. This antisera was used to confirm ~116 subcellular distribution, which appeared to be mainly particulate, and to confirm the increase of pl16 in Zucker rat obesity (Fig. 4)
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
“,
,
0
20
40
60 Time,
8'0
100120'140
.?, .;i;dr,;~rl'., .:;.~:.y: :.; .,:I:...,-, .:
carboxylase
in genetic obesity
CHRISTOPHER J. LYNCH, KENNETH M. MCCALL, MELVIN L. BILLINGSLEY, LIZABETH M. BOHLEN, STANLEY P. HRENIUK, LOUIS F. MARTIN, LEE A. WITTERS, AND SUSAN J. VANNUCCI Departments of Cellular and Molecular Physiology, Pharmacology, and Surgery, College of Medicine, The Pennsylvania State University, Hershey, Pennsylvania 17033; and Endocrine-Metabolism Division, Departments of Medicine and Biochemistry, Dartmouth Medical School, Hanover, New Hampshire 03756 Lynch, Christopher J., Kenneth M. McCall, Melvin L. Billingsley, Lizabeth M. Bohlen, Stanley P. Hreniuk, Louis F. Martin, Lee A. Witters, and Susan J. Vannucci. Pyruvate carboxylase in genetic obesity. Am. J. Physiol. 262 (Endocrinol. Metab. 25): E608E618, 1992.-Immunoblotting and protein microsequencingwere usedto identify severaladipocyte proteins expressedin an obesity-related fashion in the Zucker rat. One of these was a 116-kDa particulate protein (~116). The ~116 levelsin adipocytes from 5 to 7-wk-old obese Zucker rats were two- to fivefold higher on a per milligram of protein basisthan levelsin lean animalsand decreasedafter the induction of streptozotocin-induced diabetesmellitus. This suggests the change may be related to the actions of insulin. Hepatic levelsof ~116 did not change.The ~116 waspurified to homogeneity from obese Zucker rat adipocytes, and polyclonal antisera were prepared against the purified protein in rabbits. Microanalysis of electroblotted pII6 proteolytic fragmentssuggestedthat ~116waspyruvate carboxylase(PC). Other evidence that ~116 was PC included the following: 1) ~116 contained biotin, 2) ~116 in particulate subcellularfractions was soluble after freeze-lysis, 3) antibodies to ~116 reacted with purified hepatic PC, 4) ~116 and purified hepatic PC had identical p1 and relative molecular weight values, and 5) similar changes weredetectedin adipocyte ~116 and PC enzyme activity during obesity and after the induction of streptozotocin-induced diabetesmellitus. IncreasedadiposetissuePC probably contributes to the increasedlipogenic capacity of young obeseZucker rat adipocytes. fatty acid synthetase; adenosine5’-triphosphate-citrate lyase; carbonic anhydrase III; coenzyme A carboxylase; Zucker rat; diabetesmellitus; lipogenesis;avidin; insulin; microsequencing
betweenhumanobesity, presumed to have some genetic component, and the characteristics of Zucker fa/fa obese rats. It is not surprising, therefore, that these animals have been extensively studied regarding the biochemical, physiological, and behavioral manifestations of genetic obesity (for reviews see Refs. 3, 9, 16, 22). Important information has been gained from these studies on the biochemical changes in the adipocyte and other tissues during the development and maintenance of the obese state. Despite this, a complete picture of the disease etiology is not yet available, especially regarding the identification of the fa gene or genes. To address this problem, our laboratory is mapping and identifying specific proteins differentially expressed during obesity in adipose and
THEREAREMANYSIMILARITIES
other tissues of genetically obese Zucker rats.
In this communication, we confirm the increased expression of the known lipogenic enzymes, fatty acid synthetase, ATP-citrate lyase, and acetyl-CoA carboxylase previously reported in this model. We also report on the purification and identification of a protein whose expression increases two- to fivefold in adipocytes from E608
0193-1849/92
young obese Zucker rats and decreases during strepto-
zotocin-induced diabetes mellitus in these animals. The data suggest that this protein is pyruvate carboxylase (PC). Given the critical role of PC in maintaining tricarboxylic acid cycle intermediates (14, 18, 27, 32-33), which are presumably being actively drawn off in obesity for glycerogenesis and lipogenesis, we propose that PC contributes to the hypertrophy of adipose tissue during the development of obesity. EXPERIMENTAL PROCEDURES Isolation of adipocytes, subcellular fractions, and liver homogenates. Intact adipocyteswerepreparedfrom the parame-
trial fat pads of anesthetized and fasted lean and obesefemale Zucker rats as previously describedin detail (41). Isolated fat cellswere processedeither for gel electrophoresisor subcellular fractionation. For gel electrophoresis,isolated fat cells were washedin Krebs-Ringer bicarbonate buffer (pH 7.4) containing 200 nM adenosinebut without the usual4% bovine serumalbumin fraction V. These concentrated cellswere frozen at -85°C as 1503.Jaliquots and were assayedfor protein content using Pierce CoomassieBlue reagent. For subcellular fractionation, cells were resuspendedin 10 vol of ice-cold homogenization buffer (IO mM Na2P04, 5 mM EDTA, pH 7.4, including 5 pg/ml leupeptin, 5 pg/ml soybean trypsin inhibitor, and 5 pg/ml aprotinin). The cells were homogenized at 4°C in a PotterElvehjem homogenizerand centrifuged at 10,000g to isolate a low-speedparticulate subcellular fraction containing plasma membranevesicles,mitochondria, and nuclei. These fractions were washed during a second 10,000g centrifugation, resuspendedin homogenizationbuffer, assayedfor protein asabove, and stored at -85°C. The 10,000-gsupernatant wascentrifuged at 100,000g for 1 h to isolatethe high-speedparticulate subcellular fraction, which wasthen washedin homogenizationbuffer for 1 h at 100,000g. The 100,000-gpelletswere similarly assayed for protein content and stored at -85°C in homogenization buffer. The 100,000-gsupernatant, termed the cytosolic fraction, wasconcentrated without significantly increasingsalt levelsby vacuum dialysis against 4 liters of homogenizationbuffer in a Bio-Molecular Dynamics (Beaverton, OR) Micro-ProDiCon device. Aliquots of the concentrated cytosolic fraction were assayedfor protein, and the remainder was frozen at -85°C. Livers from these animalswere briefly perfusedand homogenized in ice-cold homogenizationbuffer using a Tekmar tissue homogenizerand stored in aliquots at -85°C. Gel electrophoresis. One-dimensionalsodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and Coomassie Blue staining were performed as previously described (21) using a Hoefer SE600 apparatus with 4°C cooling. Twodimensional isoelectric focusing (IEF)-SDS-PAGE and silver staining weredoneaccordingto Hochstrasseret al. (19). Briefly, adipocyte or liver proteins were diluted to 1.5 mg/ml in 10% SDS and 23 mg/ml dithiothreitol. These mixtures were solubilized using a Tekmar Vibracell sonicatorfitted with a 2-mm microtip probe and mixed in a 1:4 ratio with a solution contain-
$2.00 Copyright 0 1992 the American Physiological
Society
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
PROTEIN PROTEI N EXPRESSION
1
2
IN GENETIC
34
56
116~
Fig. 1. Adipocyte proteins from lean and obese Zucker rats. Adipocytes (70 kg protein/lane) and adipocyte low-speed particulate subcellular fractions (50 fig protein/lane) prepared from parametrial fat pads of 5- to 7-wk-old heterozygous lean and homozvsous obese Zucker rats. as described in text, were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) and Coomassie Blue staining. Lane 1, adipocytes from lean rats; lane 2, adipocytes from obese rats; lanes3 and 4, lowspeed particulate fractions from lean rats; lanes 5 and 6, low-speed particulate fractions from obese rat adipocytes. Arrow indicates 116-kDa protein band (~116) M,, relative molecular weight.
67-
cy) 10 l-
E609
OBESITY
“-
43-
X
nrlinnnzrtn
dye and waselectrophoresedat per gel until the tracking dye migrated to the bottom of the Zucker ruts as compared with lean lit(term&es resolving gel. Protein Concentration in Obese flm3ntification Purification of ~116, antibody production, and Western Designation Adipocytes Relative to Lean blotting. Polyclonal antibodies to ~116 were preparedin rabbits ~265 (C) elevated (19-fold higher) Acetyl-CoA carboxylase* as follows. The ~116 was purified from 10,000-gparticulate ~260 (C) elevated (19-fold higher) Fatty acid synthetase* fractions of obeseZucker rat adipocytes using preparative (100 ~116 (P) elevated (2- to 5-fold higher) Pyruvate carboxylase rugprotein/tube) two-dimensional IEF-SDS-PAGE. The ~116 ~112 (C) elevated (19-fold higher) ATP-citrate lyase* ~106 (P) decreased (80-90% less) Unknown gelspot centerswereremovedfrom CoomassieBlue-stainedgels decreased (80-90% less) Unknown in destain usinga 3-mm coring tool. Twelve of thesespotswere decreased (40-60% less) Carbonic anhydrase III? collectedper iniection. Gel snotswere eouilibrated for 10 min in Seven major proteins (p) were initially observed to be differentially 10ml of water,placed in microfugetube;, frozen in liquid nitroexpressed in a series of one-dimensional gel electrophoresis experiments gen, and lyophilized in a Jouvan RC 10.10 centrifugal freeze such as that shown in Fig. 1. P, particulate fraction; C, cytosolic dryer. Dried gel spots were powdered with a microfuge tube fraction. See text for explanation of pyruvate carboxylase. * Identified homogenizing pestle and resuspendedin 1.5 ml of sterile isoin immunoblot studies (not shown) using previously described antisera (6, 12). Changes in acetyl-CoA carboxylase mass were also confirmed tonic phosphate-bufferedsaline. This gel mixture wasdrawn up by enzyme-linked immunosorbent assay, as previously described (20). into a 5-ml luer-lock glasssyringe containing 1.5 ml of either tconfirmed by microsequencing isoform-specific antisera and enzy- Freund’s completeadjuvant (Sigma), Freund’s incompleteadjumetic activity measurements (C. J. Lynch, K. M. McCall, R. L. vant, or 8 mg/ml of aluminum hydroxide adjuvant. These mixHoretsky, S. J. Vanucci, N. Carter, and S. J. Dodgson, unpublished tures wereemulsifiedusing a stainlesssteelmale-maleleur-lock observations). microemulsifying tube (ZO-guageX 7/8 in.; Thomas Scientific) ing 4% 3-[(3-cholamidopropyl)dimethylammonio]-l-propane- connected to another 5-ml leur-lock tube. The mixture was sulfonate detergent, 54% urea, 10 mg/ml dithiothreitol, and 4% injected subcutaneously and intradermally at multiple sites ampholines[a 4:l mixture of BDH Chemicals(Poole, UK) pH along the shavedback of a male New Zealandrabbit (~116with 3-10 and pH 5-7 40% ampholine solutions]. This mixture was Freund’s adjuvant) or intraperitoneally (~116 with aluminum sonicated, and 50 ~1 were applied to 120 x 1.5 mm IEF poly- hydroxide adjuvant). The injection, adjuvant, and bleeding acrylamide (4% acrylamide) tube gelsusing diacrylylpiperazine schedulewasthat of Green et al. (15). For humanepurposesand (BioRad) asa cross-linker instead of N,N’-methylene-bis-acry- as a bleedingaid, rabbits were anesthetizedusing a Droperidollamide (19). Sampleswere electrofocusedin a Hoefer tube gel Fentanyl mixture (Innovar-Vet) during every procedure.Immuapparatusfor 7,500 V hours at room temperature according to noblotting was performed according to Lynch et al. (26) using the following program: 200V for 2 h, 400 V for 15 h, and 800 V i”“I-labeled goat anti-rabbit immunoglobulin G F’Ab, fragfor the remainder.Tube aelswereeentlv extruded usinadistilled ments and autoradiography as a meansfor identifving antibodv water, blotted briefly without touihingthe gel,and gently rolled binding proteins or, as describedby Bianchi et al: (6r, for fatty between two glass plates filled with SDS-PAGE resolving acid synthetase and acetyl-CoA carboxylase. For detection of buffer, the 4% stacking and 7.5% resolving SDS-PAGE. The biotin-containing proteins on Western blots, we used avidintube gel in contact with the 4% gel was covered with 1 ml of linked alkaline phosphatase,as describedby Billingsley et al. Laemmli sample buffer (21), containing bromophenol blue (7). Tahlo 1_cIUIL.
1I.
Phnnnnc “rYLMv~6co
;n “I”
mninr rr”uJ”r
ULufJ”LJY~,
vrot&s
in
,-&se
tracking
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
E610
PROTEIN
EXPRESSION
IN GENETIC
IEF-PAGE
OBESITY
IEF-PAGE
OBESE
LEAN ADIPOCYTES
-116 -97.4 -67
:
a a I
-43
iz co
-29
OBESE
LEAN LIVER
Fig. 2. Proteins from adipocyte lysates and liver homogenates. Adipocyte lysates (17 wg protein/tube gel) and liver homogenates (17 pg protein/tube) subjected to two-dimensional isoelectric focusing (IEF)-SDS-PAGE and silver staining. Top left: adipocyte lysate from lean rats. Top right: adipocyte lysate from obese rats. Bottom Left: liver homogenate from lean rats. Bottom right: liver homogenate from obese rats. Arrow indicates ~116 (p1 = 7.0, size = 116,000 kDa).
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
PROTEIN
EXPRESSION
IEF-PAGE
IN GENETIC
OBESITY
E611
100KCl, 50 tris(hydroxymethyl)aminomethane.HCl (pH 8.1 at 37”C), 15 [14C]KHC03, 2.4 ATP, 0.2 acetyl-CoA, 5 MgC&, 0.1 ethylene glycol-bis(&aminoethyl ether)-N,N,N’,N’-tetraacetic acid, 1 NADH, and 65 U of malate dehydrogenase(Boehringer Mannheim; “special quality”) in the presenceor absenceof 10 mM sodiumpyruvate in a total volume of 1 ml. Reactionswere stopped with the injection of 1 ml of 2 N sodium acetate (pH 3.4). This mixture wasshakenat 37°C for 1 h. After that period, the center well containing hyamine hydrochloride asa COstrap was removed, and the radioactive content of the reaction mixtures was determined using a Beckman LS-3801 liquid scintillation counter. PC waspurified from rat liver using the method of Ahmad et al. (1). The specificactivity of the purified enzyme was31 U/mg protein and was stored (-84°C) at a concentration of -1.0 mg protein/ml. RESULTS
Fig. 3. Low-speed particulate adipocyte proteins from obese rat. Silver stain of two-dimensional IEF-SDS-PAGE of low-speed particulate subcellular fraction from obese Zucker rat adipocytes.. Such gels subjected to reversible Coomassie Blue staining were used as source of ~116 (arrow). Internal amino acid sequence analysis of ~116. Limited proteolysis of purified p 116 by the Clevelandet al. (11) method was obtained when’ Staphylococcus aureus V8 protease (l-2 ng protein/gel lane) and ~116 (lo-20 Kgprotein/lane in gel circles isolatedasdescribedabove) were allowedto electrophoreseinto a 4% acrylamide stacking gel poured over a 12% acrylamide SDS-PAGE resolving gel and then allowedto incubate at 30°C for 30 min without current. After the digestion period, electrophoresis resumedat 25 mA constant current, and the buffer chamber was cooled back to 4°C. When the tracking dye reached the bottom of the 12% acrylamide SDS-PAGE, peptides were transferred to polyvinylidene difluoride (PVDF) membranesat 500 mA constant current for 90 min in transfer buffer containing 10 mM 3-(cyclohexylamino)-l-propanesulfonic acid (pH 11) and 10% methanol, asdescribedby Matsudaira(29). The blotting apparatus (BioRad) contained a cooling core through which 4°C coolant was circulated. Visualization of proteins and peptideswasby CoomassieBlue staining (29). Bands containing peptides were cut with a clean razor blade, and sequenceanalysis was performed using an Applied Biosystemsmodel470-A sequenatorequippedwith on-line phenylthioisocyanate analysis, as describedby Matsudaira (29). PC activity and purification. PC activity was assayedunder conditions needed for maximal activity, as described by McClure et al. (30-31) using a radioisotopic method for the detection of pyruvate-dependent [ 14C]bicarbonatefixation as follows: 25 ~1of samplewere incubated at 37°C with (in mM)
In initial one-dimensional gel electrophoresis experiments, such as those depicted in Fig. 1, we reproducibly observed differential staining of seven major proteins from adipocytes of obese Zucker rats compared with lean littermates on a per milligram protein basis (Table 1). To identify these proteins, subcellular fractionation studies using isopycnic centrifugation (Fig. 1 and data not shown) and metabolic radiolabeling experiments employing inorganic [32P]PO: were initially performed (C. J. Lynch, K. M. McCall, R. L. Horetsky, S. J. Vanucci, N. Carter, and S. J. Dodgson, unpublished observations). This data together with the pI and relative molecular weight (M,) of the proteins, as determined by us, was used to computer search adipocyte, Zucker rat, and 3T3 cell line literature. From these efforts the identities of three of the cytosolic (i.e., 100,000-g supernatant) proteins with relative molecular sizes of -265, 260, and 112 kDa, were suggested, i.e., acetyl-CoA carboxylase (p265), fatty acid synthetase (p260), and ATP-citrate lyase (~112). This was confirmed by quantitative immunoblotting studies using antisera that specifically recognized acetyl-CoA carboxylase and antisera that recognized both fatty acid synthetase and ATP-citrate lyase (Table 1). The identification of ~112 in our gels was further substantiated in other experiments (see below) with ATP-citrate lyasespecific antibodies kindly supplied by Dr. Charles Rubin (Albert Einstein College of Medicine, New York, NY). The identity of the other four proteins could not be so determined. These included three particulate proteins of 116, 106, and 88 kDa and a 28-kDa cytosolic protein, termed ~116, ~106, ~88, and ~28, respectively. Three of the proteins, ~106, ~88, and ~28, displayed decreased staining when SDS-PAGE of adipocyte proteins from obese Zucker rats was compared with lean littermates (Fig. 1 and Table 1). The ~28, an abundant protein in lean rodent adipose tissue, has been recently purified and identified by microsequencing as well as two-dimensional immunoblotting as carbonic anhydrase III (C. J. Lynch, K. M. McCall, R. L. Horetsky, S. J. Vanucci, N. Carter, and S. J. Dodgson, unpublished observations). Purification and identification of p106 and p88 is now proceeding in our laboratory. Staining of the other, a li6-kDa particulate protein (~116, the primary subject of the present study), was increased two- to fivefold on a per milligram protein basis in adipocytes from obese Zucker rats (Fig.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
rats were subjected to subcellular fractionation (50 pg protein/lane) and SDS-PAGE. Proteins were transferred from gels to Nitroplus transfer membrane (Micron Separations, Westboro, MA) and subjected to immunoblot analysis using lz51-labeled anti-rabbit gamma-globulin F’Ab2 fragments followed by autoradiography of resulting blots. Relevant molecular weight region of immunoblots from duplicate gels. Lane I, whole cell lysate; lane 2, low-speed particulate subcellular fraction; lane 3, high-speed particulate subcellular fraction; and lane 4, concentrated high-speed supernatant. A: antisera was raised against purified ~116 (1:200 dilution). B: antisera was raised against purified ATP-citrate lyase (1:2000 dilution) and was a generous gift of Dr. Charles Rubin, Albert Einstein College of Medicine, New York, NY.
Table 2. Amino acid alignment of internal sequences from ~116 with yeast PC Item
Species
Homology with p116, %
Aligned Sequence
24K peptide
Rat
LFQLXPAQNRAQKLL
23K peptide
Rat
LFQLRP AQNRAQKLLHYLG
PC
Yeast
20K peptide
Rat
PC
Yeast
18K peptide
Rat
PC
Yeast
lOT
70
lOT
LFQMV: ZSQNRAQKLLHYLA 480T
AAISYXl SXVADPSR
lOT
25
ALKDVI (GQIGAPMP 301 KIGINGXAIQXRVQAEXPA
63
KITTRGFAIQCRITTEDPA
lOT
360T
* Purified ~116 was subjected to limited proteolysis using Stuphylococcus aureus V8 protease. Several peptides on polyvinyl difluoride Western blots resulting from this digestion were subjected to automated gas-phase sequencing as described in text. They are compared with databank sequences submitted by Lim et al. (25) and deduced from a cloned fragment of genomic DNA containing pyruvate carboxylase yeast (PC) gene. Percent homology refers to % exact matches excluding Xs (unrecognizable or ambiguous amino acid in 2 separate sequencing attempts). Decapeptide marks appear below sequences.
1). In two-dimensional analytic gels (17 pg protein/tube), a single 116-kDa protein could be observed, which had a p1 of 7.0 and displayed increased staining in gels from obese relative to lean animals (Fig. 2, top). A protein with identical mobility in two-dimensional gels was also seen in livers from Zucker rats (Fig. 2, bottom). The abundance of the 116-kDa liver protein was not influenced by obesity, however. The ~116 purification
and immunoblotting studies. We
purified ~116 by preparative two-dimensional IEF-SDSPAGE for polyclonal antisera production and protein sequencing. Because ~116 seemed to be most abundant in low-speed particulate fractions from obese Zucker rats, as determined from one-dimensional SDS-PAGE and Coo-
massie Blue staining (data not shown), this preparation was used as starting material for purification by preparative two-dimensional IEF-SDS-PAGE. Figure 3 shows a representative preparative two-dimensional IEF-SDSPAGE of adipocyte particulate proteins from obese Zucker rats. The most abundant protein in such gels was P 116 (i.e., p1 = 7.0, M, = 116,0&O), as detected by CcIOmassie Blue (data not shown) or silver staining (Fig. 3). The centers of such spots from Coomassie Blue-stained gels, which should theoretically contain only pure ~116 protein, were used to prepare polyclonal antibodies in rabbits and for microsequencing studies. The antisera were used in two-dimensional IEF-SDS-PAGE immunoblotting experiments to show that the protein identified in liver was ~116. Figure 4 shows that the immunoreactivity of our antisera is against ~116, since the 116-kDa immunoreactivity is most abundant in adipocyte particulate fractions and is increased fivefold in Zucker rat obesity, as opposed to ~112 (ATP-citrate lyase) immunoreactivity, which is localized in the cytosolic fractions and is increased -19-fold in fractions from obese relative to lean animals.
Studies to determine the identity of ~116. Coomassie Blue-stained PVDF blots were obtained from ~116 digests using Staphylococcus aureus V8 protease for limited digestion and SDS-PAGE containing 12% acrylamide for peptide separation. Such digestions yielded four peptides that did not comigrate with peptides found in lanes with protease alone (data not shown). The peptides had relative molecular sizes of 24, 23, 20, and 18 kDa. These were subjected to gas-phase microsequencing in two separate experiments using different batches of protease and ~116 (Table 2). The 24- and 23-kDa peptides appeared identical at their NHs-termini, with the exception that, in two separate experiments, ambiguous sequence information was obtained from the 24-kDa peptide after the fifteenth sequencing cycle. The alignment routine of the DNA-STAR computer program was used to search the National Biomedical Research Foundation Protein Identification Resource for matching amino acid sequences. The best overall homology was with PC from
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
PROTEIN
EXPRESSION
IN GENETIC
IEF-PAGE
12
W (3 a a A CI co
v-”\ h--i : ‘j/ 1 I _i
.,‘
/*
/ : -_ >r
E613
OBESITY
/ - I/’_ .‘;Ci’2/(
\-i3-%a - I I i
34
?/iYi ’ \ 1tI \ / 51 : iI i / \** 1; L/ // jj is i ;? 3: /’ : \i I x : i i \ /i n * * 3* , , -\ \\ _ tt 3 , ’ I ,I1’r*I /
Fig. 5. Avidin binding proteins in adipocytes and liver. Proteins on Western blots binding avidin conjugated to biotinylated alkaline phosphatase were visualized using nitro blue tetrazolium chloride-5-bromo-4-chloro-3-indoyl phosphatep-toluidine salt chromagen system (7). Left: low-speed particulate subcellular fractions from obese Zucker rat adipocytes were separated by two-dimensional IEF-SDS-PAGE. Arrow indicates position of ~116 (compare with stained gel in Fig. 3). Right: SDS-PAGE of adipocyte lysates from lean (lane I) and obese (lane 2) Zucker rats or liver homogenates from lean (lane 3) and obese (lane 4) Zucker rats.
yeast (25), having between 25 and 70% exact matches with internal peptides. To compare our sequences with those of full-length mammalian PC sequences, several other data banks (i.e. Swiss Protein data bank release 15, European Molecular Biology Laboratories data bank release 24, and GenBank release 65) were searched using either the term PC or EC 6.4.1.1. These searches did not reveal published full-length mammalian cDNA encoding PC which could be translated. The homotetrameric subunits of yeast PC each have a calculated molecular mass of 116 kDa and a calculated p1 of 7.0, i.e., like ~116 (25). However, we were not entirely convinced that ~116 was PC in the absence of a translatable full-length mammalian cDNA. One reason for contern was that earlier studies on mammalian PC reported a relative molecular mass of - 130 kDa (e.g., Refs. 13, 14, 31), somewhat higher than ~116. Several added experiments were performed to test the hypothesis that ~116 was PC. PC contains a biotin moiety that transfers bicarbonate to form oxaloacetate. The protein avidin tightly binds to biotin residues (7), and avidin conjugated to a chromogen-generating system was used to look for highaffinity avidin binding by ~116, as described in EXPERI-
PROCEDURES. Figure 5 (left> shows that ~116 isolated in two-dimensional IEF-SDS-PAGE and immobilized on nitrocellulose membranes does bind an avidin conjugate. This 116-kDa avidin binding activity was less abundant in liver homogenates compared with adipocytes and was increased in adipocytes from obese relative to lean Zucker rats (Fig. 5, right). This suggests that activity represented binding to biotin attached to ~116 and not binding to some other biotin-containing 116-kDa protein, PC is a soluble protein but is found almost exclusively inside mitochondria and therefore would not be expected to be found in cytosolic fractions (3). Manipulations that break mitochondria release a soluble form of PC. The ~116 was found in low-speed particulate subcellular fractions containing mitochondria but not cytosol (Fig. 4). However, it was unclear whether ~116 was a soluble protein, an integral membrane, or cytoskeletal protein, Figure 6 shows that, when freeze-lysed 10,000-g pellets were sonicated in hypotonic buffer to turn mitochondria inside out, ~116 was released in a soluble form which would not sediment at 100,000 g. Figure 7 shows that ~116 antibodies react with PC purified from rat liver, and Fig. 8 shows that purified MENTAL
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
E614
PROTEIN
EXPRESSION
IN GENETIC
OBESITY
1
1
-pi /,
16
Fig. 6. Release of soluble ~116 from low-speed particulate subcellular fractions after freeze-lysing and sonication in hypotonic media. Freshly prepared low-speed particulate subcellular fractions from obese Zucker rats, diluted to 1 ml (1 mg protein/ml), were frozen in liquid Ns, thawed, and sonicated. This suspension (50 ~1) was reserved (lane 1) and remainder was centrifuged at 100,000 g for 1 h. Supernatant (50 ~1) was removed (lane 2), and both samples were subjected to SDS-PAGE followed by Coomassie Blue staining.
hepatic PC and ~116 comigrate in two-dimensional IEF-SDS-PAGE. Last, Fig. 9 shows that PC activity was increased in young obese Zucker rat adipocytes when these were compared with adipocytes from lean controls. The increase in activity was consistent with the increases in ~116 immunoreactivity, avidin binding, and staining, which ranged from two- to fivefold in our studies. All of the above findings support the hypothesis that ~116 is PC. Effect of streptozotocin-induced diabetes mellitus on PC. Caro et al. (10) have forwarded a model of human obesity in which hyperinsulinemia and insulin resistance, progressing with age more slowly in adipose tissue vis-a-vis other peripheral tissues, result in preferential shunting of substrates to adipose tissue resulting in adipocyte hypertrophy and adipose tissue hyperplasia. Consistent with Caro et al.3 model, young obese Zucker rats display hyperinsulinemia and increased adipose tissue responsiveness to insulin (9, 17). To determine the possible role of hyperinsulinemia in the obesity-related increases in PC that we observed, the effects of streptozotocin-induced diabetes mellitus on adipocyte PC concentrations and activities in the obese Zucker rat were studied (Figs. 9-10). Figure 9 shows that acute diabetes mellitus reversed the effects of obesity on PC activity in
Fig. I. Recognition of uepa~~ pyruva~ cdruoay~x antibodies. Purified hepatic PC was subjected to SDS-I m-r!, on IWO acrylamide gel. One-half of gel containing 2 pg of purified hepatic PC was Coomassie Blue stained (lane 1). Other half (lane 2) also containing 2 pg of purified hepatic PC was transferred to Nitroplus transfer membrane and subjected to immunoblot analysis using ~116 antibodies, as
describedin Fig’ 4’ adipocytes. Figure 10 shows that concentrations of PC (~116) decreased in acutely diabetic obese animals. Concentrations of ~265, ~260, and ~112 also decreased in diabetic obese rats while p28 increased. The ~106 and ~88, major proteins in adipocytes from lean animals (Fig. l), were not readily observed in Coomassie Blue-stained SDS-PAGE of either obese or diabetic obese rat adipocyte proteins (Fig. 10). These findings support the hypothesis that PC mass and activity increase in adipocytes from young obese Zucker rats and that such changes may be related to the actions of insulin. DISCUSSION Seven major adipocyte proteins and phosphoproteins were initially observed to be altered in obesity (Fig. 1 and
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
? co
IEF-PAGE
I
C
Fig. 8. Comigration of purified hepatic PC and adipocyte ~116 in twodimensional IEF-SDS-PAGE. Proteins from low-speed particulate subcellular fraction from obese Zucker rat adipocytes containing ~116, indicated by arrow (A), purified hepatic PC (B), or 1:l mixture of both (C) were subjected to two-dimensional IEF-SDS-PAGE and silver staining. STD, lane of standard molecular weight markers.
! _ “/
” i*buer*w*n-e - NUwrVI_
Table 1). From previous studies on adipocytes or 3T3 cell lines (e.g., 5, 8, 23, 28, 36-39), we deduced and subsequently confirmed the identity of three of these by immunoblotting to be acetyl-CoA carboxylase (p265), fatty acid synthetase (p260), and ATP-citrate lyase (~112) (Table 1, Fig. 4, and data not shown). The direction of the change in the expression of these proteins in the obese Zucker rats is in keeping with previous studies on their activity in this animal model (5, 23, 28, 37, 39). However, the magnitude of the change determined in this study was higher, i.e., 19-fold. One possible explanation for this difference is that the activity of some of these enzymes is regulated by phosphorylation (34) and other mechanisms (2, 24) that may be preserved during cell breakage and enzyme assay. Our findings are the first evidence that the actual mass of these enzymes is increased in the obese Zucker rat. The identification of the other four adipocyte proteins found altered in obese Zucker rats was not so determined. Purification and identification of two particulate proteins, namely ~106 and ~88, is now under way. A
third, ~28, has been identified by protein microsequencing, immunoblotting, and other experiments as carbonic anhydrase III (C. J. Lynch, K. M. McCall, R. L. Horetsky, S. J. Vanucci, N. Carter, and S. J. Dodgson, unpublished observations). The present communication describes the purification and identification of the fourth, ~116. Coomassie Blue staining of ~116, a ll6-kDa major adipocyte protein, was increased between two- and fivefold in SDS-PAGE of adipocyte proteins from obese compared with lean Zucker rats (Figs. 1-2). A liver protein that was identical to ~116 in terms of two-dimensional IEF-SDS-PAGE mobility did not change in response to obesity (Fig. 2). The ~116 protein was purified to homogeneity by preparative two-dimensional IEF-SDS-PAGE using high-density adipocyte membranes from obese Zucker rats as starting material (Fig. 3). Polyclonal antibodies were prepared against the purified protein. This antisera was used to confirm ~116 subcellular distribution, which appeared to be mainly particulate, and to confirm the increase of pl16 in Zucker rat obesity (Fig. 4)
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.154.019.149) on January 13, 2019.
“,
,
0
20
40
60 Time,
8'0
100120'140
.?, .;i;dr,;~rl'., .:;.~:.y: :.; .,:I:...,-, .:
