THESIS This is to certify that the dissertation entitled STUDIES ON MAMMALIAN METAPHASE CHROMOSOMES AND CHARACTERIZATION OF PHOSPHOFRUCTOKINASE MEMBRANE BINDING AND ACTIVITY INHIBITION BY HEXACYANOFERRATE(ll) presented by David Phillip Lapenson has been accepted towards fulfillment of the requirements for _Ph-D—odegree in BIOCHEMISTRY m Damwrfla MSUI‘HIH‘” ,- . - r1 'nrr - . . . 0-12771 MSU LIBRARIES m \r RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. STUDIES ON MAMMALIAN METAPHASE CHROMOSOMES AND CHARACTERIZATION OF PHOSPHOFRUCTOKINASE MEMBRANE BINDING AND ACTIVITY INHIBITION BY HEXACYANOFERRATE(II) By David Phillip Lapenson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistny 1983 ABSTRACT STUDIES ON MAMMALIAN METAPHASE CHROMOSOMES ND A CHARACTERIZATION OF PHOSPHOFRUCTOKINASE MEMBRANE BINDING AND ACTIVITY INHIBITION BY HEXACYANOFERRATE(II) By David Phillip Lapenson I. Two methods, moving-boundary electrophoresis and a rapid iso- electric focusing procedure, were investigated as possible techniques for obtaining quantities of pure individual types of mammalian meta— phase chromosomes. In both cases, the experiments were hampered by precipitation of chromosomes in the electric field. Conditions were found under which precipitation could be minimized, but not eliminated. 2° At PH 6.8, pig kidney phosphofructokinase, yeast glucose-6- PhOSphate dehydrogenase, and yeast phosphoglucose isomerase were corn- Petitively inhibited 90, 30, and 30%, reSpectively, by I mfl hexacyano- ferrate(II), in the presence of 0.2 m hexose-phosphate substrate. UHIIke all previously reported inhibitions of glycolytlc enzymes by Thexacyanoferrate, this inhibition does mt involve 925.193.1231! 0f enzyme, Substrate, or enzyme-substrate complex. Rather, it appears to be due to a reversible binding of hexacyanoferrate at, or near, the hexose-phosphate binding site of the enzyme. David Phillip Lapenson 3. In differential centrifugation studies at pH 7.4, we found that 50 to 75% of the phosphofructokinase activity in pig liver homogenates is particulate, and associated with a specific membrane fraction. In sucrose density gradient equilibriwn centrifugation experiments with step—gradients, particulate phosphofructokinase floats, under condi- tions where nuclei and aggregated proteins sediment to the bottom. These studies suggest that pig liver phosphofructokinase binds to the plasma membranes. 4. The particulate phosphofructokinase is solubilized immediately by addition of millimolar concentrations of fructose-6-P, fructose- l,6-P2, glucose-6-P, or AMP, to the homogenate. When the added hexose-phosphates are depleted, the solubilized phosphofructokinase again becomes particulate. There is a direct correlation between the disappearance of solubilizer metabolites and formation of particulate phosphofructokinase. Triton X-l00 (l%) solubilizes, but relatively slowly, requiring nearly an hour for maximum solubilization. 5. In contrast, rat liver phosphofructokinase appears to be entirely soluble. Rabbit liver phosphofructokinase sediments with the microsomal fraction, but is not bound to microsomes (or glycogen), and appears to be highly self-associated. 6. In experiments with pig hepatocytes, most of the phosphofructo- kinase activity appeared to be lost during the hepatocyte isolation, and the low remaining activity was soluble. Total phosphofructokinase activity (measured at high fructose—6—P concentration) increased or David Phillip Lapenson decrwaased predictably when hepatocytes were incubated at 37° in high concentrations of glucose or lactate, respectively. VITA T)avimv use .Ammvv .ocv e__e wqoomoeowe cocpum_m xenox co umgamcmopoga mew; mommeH .>x on Co mmmp_o> mcwpmcmFmooe cm cue: .xooo.m¢ ow.xooo~ eo mcowpmoww_:mme we .2me Fom ma_Fw;a m :_ nozmw> mew; muwew .muogpmz cw vwnweommv we cm_m .pa sew: uwZOGmcm vce .muwem emmeUYcowuo_ema :0 a: umx0wa .mpmeuw: Pacmez c? umcwmpm mew: masomoeoeco .muogumz use mFewcmpmz cw umnwcommu we .ANNV e_memm—nnzpm new Ame: new .ANV wuempy< ego :meLwnsz Co meocpwe mg» we :o_peuwewuoe m he ummeOmw mew; meOmoeoegu mmmcampmz .soe>sw mzms>xmz.z _ ZH w.e Ia e< omem<>o mmemz4w mzm4>xmz.8 F 2H .w.o Ia H< amH5 anecwmuno covpecowpoeee wN_m Feowqap ecu bcmmmeame czogm chVpoecw emcee .x com u cowpmowmwcmee FepOH .omio cw umao_m>mu use .Aoo¢-wexucmzp .muocpmz vcm m_mwcmpmz cw nonweommv mm .pcmwumem mmoeozm Roe ow o_ w :o umbmcowpomew mew; _ooxpm mcmfixxm;.z _ cw 2a chgzm: be nmmeomw mesomoeoeco mmmcqmpmz .onHHHuoow> HzmHom omzH Is ms» .umpuowusw mu .Ls NP so Ls s so; msoowosuomrw op uwZOP—u .mpsmewsmsxm —mfi_usus oz» Eose mum szosm mwfiweoss mosussomsu use .Am.u._v muospwz usu mpuwswuuz s_ umswsomwu mu .wscwssomu m—ssumiwp_se .uwsus wsp mswm: ummsooeospomFm use ._ooa_m msmFxst.a F +0 wosmmwss use sw umquOmw mew: mweomoeossu .mssmo ozu 20mm mmZOmozoqu mm_uoe esp .sowpespswosou seem u< ._e\m: s.o we; sowuespseosoo yes esp .sioiwmouusse.ze u msws_epsoo useewsesxe esp sH ._e\ms s me: so_pesuseosoe gas use .sieiemoposee we msovpespseosoo as m.o use .N.c ._.o mswsweesoo mesmewcesxe esu sH .uems me: esspxwe sowpoees Mus ugeusepm use .muosuez use m~e_empez sw ueswsomeu we .siuiemopoese use AHHVepessmeoseAQexes eo msowuesuseosoe meowse> we musemess es» s? uexemme we; ape>wpue yes .AHvaHom Axsmv mmmonx GHQ so onFHmszH mzh zo Asmsv simimmOHusmm so onHm<> do Humsmm .m mmast 50 Qfim cm N mesmws as . D. 35.53252?“— .~ cc; Qfic J J u q q ’ U I! It/u as 2:. 2 \es 2.: as l./. cm AIIMIOV )Hd % 51 activity was still 10- to 100- fold in excess, even in the presence of 2_pM_hexacyanoferrate. Furthermore, the extinction coefficient Of NADH at 340 nm was not affected by hexacyanoferrate. This control was important, since it would have appeared that the enzyme was inhibited if the extinction coefficient of the NADH at 340 nm had been decreased in the presence of hexacyanoferrate. 3. Oxidation State of Hexacyanoferrate In the Assay Since the inhibition experiments were conducted by adding hexacyan- oferrate in the oxidized (III) form to assay mixtures which contained a high concentration of reducing agent (50 mM 2-mercaptoethanol), it was necessary to determine the state of oxidation of hexacyanoferrate in such assays. In the absence Of 2-mercaptoethanol, the absorbance Spectrum of hexacyanoferrate in the (III) state shows three major peaks, at wavelengths of approximately 270, 320, and 420 nm respectively; shoulders are apparent on both sides Of the 320 nm peak (Figure 8). A 390 nn shoulder is apparent on the 420 nm peak. The addition Of an equimolar quantity of 2-mercaptoethanol to a solution contianing 0.5 mfl_hexacyanoferrate(III) largely abolished the two absorbance peaks at the higher wavelengths. These two peaks are completely abolished in the presence Of 1 mM 2-mercaptoethanol, so reduction is complete with only two 2-mercaptoethanol molecules per hexacyanoferrate molecule. Since the PFK assay contains a 50-fold excess of 2-mercaptoethanol, the hexacyanoferrate must be totally in the reduced (II) form. 52 FIGURE 8. SPECTRAL ANALYSIS OF THE REDUCTION OF HEXACYANOFERRATE(III) BY 2-MERCAPTOETHANOL. Varying concentrations Of 2-mercaptoethanol were added to solutions originally containing 0.5 mM_potassium hexacyanoferrate(III), K3Fe(CN)6. The sample cuvettes also contained standard assay buffer with 0 to 50 mM 2-mercaptoethanol (shown by arrows). The reference cuvettes were identical except they had no hexacyanoferrate. Spectra were taken on a Cary 15 recording spectrophotometer using 50 nm per chart division. The tungsten lamp was used to scan from 500 to 350 nm; the hydrogen lamp was used to scan below 350 nm. AB SORBANCE 53 "‘ 0.5 mM HEXACYANOFER RATE 11 . \\ ’3. :5 .- a“? .-' 1 '= [2-MERCAPTOETHANOL] f. '1 '1... <—o mM ‘ l ,. —O.1 mM \\ YE... j/O’.‘o 2.. ’0'... lo \ ‘3... .....\ Fir-4. 25 mM\.:"-..‘ 50mM——> V40. 5mM\\a"- ~—1 mM I l A . \ 250 300 350 400 450 WAVELENGTH , nm Figure 8 54 4. Proof That the Inhibition Is Not Due to Competition of Hexacyanoferrate(II) With Mgz:. Increasing the concentration of Mg2+ in the assay did not relieve the inhibition (Figure 9), and therefore hexacyanoferrate(II) must not compete for M92+. 5. Test for Competitive Inhibition With Respect to Fructose-6-P. The previous results (Figure 7) suggested that the inhibition of PFK by hexacyanoferrate(II) might be competitive with respect to fructose-6-P. The results fron a detailed kinetic study are consistent with this possibility (Figure 10A). The lines on the Linewever-Burk plot approach approximately the same Vmax value. However, the lines intersect slightly to the right of the ordinate, which might indicate an increase in Vmax in the presence Of hexacyanoferrate and high concentrations of fructose—6-P substrate. Further studies showed that hexacyanoferrate(II) did not increase the PFK Vmax’ even at concentrations of fructose-6-P above 5 mM, so this apparent intercept to the right of the ordinate was assumed to be due to experimental error. A K1 value Of 0.15 i 0.03 mM was calculated for both concentrations of hexacyanoferrate. These overall results are consistent with hexacyanoferrate(II) reversibly bindng at, or near, the catalytic fructose-6-P binding site to produce the inhibition. The v vs. s graph in Figure 108 shows the hexacyanoferrate(II) ef- fects on PFK at low substrate concentrations, where allosteric effects are Observed. The sigmoidicity of the curves become more pronounced as the hexacyanoferrate(II) concentration is increased; also, the maximal velocity appears to be lower. Both Of these results would be produced 55 FIGURE 9. EFFECT OF MgC12 ON THE INHIBITION OF PIG KIDNEY PHOSPHOFRUCTOKINASE BY HEXACYANOFERRATE(II). PFK was assayed in the presence and absence of 0.5 mM hexacyano— ferrate(II) as described in the legend to Figure 7, with the modifica- tion that additional amounts of MgCl2 were added to the assays to give the final concentrations shown on the graph. The concentration of PFK was 4 pg/ml. RELATIVE PFK ACTIVITY 56. ' 240 - , P. 200 - , N0 HEXACYANOFERRATE(I 1) .. o. ooooooooooooo o..........._° 160 - 120 ~— 30 L- o.5 mM HEXACYANOFERRATE(I i) I- O 5 ur— 40! I l l L 5 10 15 20 25 [MgCIz] , mM Figure 9 57 .Awuv sowswawz we uospee esp uem: sues; Eesmoss LewesEoo e An uesPELepeu wee; <0? essmwu see mpseosmpsw use mesofim ms» .N mesmeu op usmmep esp sw umewsommu me uexemme we; _E\ms u we soepespsmosou e we xss .AHva5um mmmonx st so onHHmHIzH so zmHzum mmzmo mh so onHHmHIzH .F— mmszs 62 ._ messes as . Z sezsssszfiesea 53185552..“ 2... cm / moo SE Nd co cw AllAllOV 0d90 lNHOHI—Id 63 .25 N.o +0 corpMchmocoo m op ¢-mummoposgw mcwuum >5 Umpgmpm mm: zwmmm mcp .XprmLmawm mmupw>so asp op umuum mm: Focmgpwogamogmsum wasp cow» -mowwwvoe mg“ cpwz muosgmz ucm mewmemz cw umnwgommu mm vmxmmmm mm; _E\m: OP mo cowpmgpcmocoo a pm Hog .AOzom mm no onpHmszH .N— mmonu 64 N_ «gnaw; as . Tzétoziogé m w m N fl c _ a d q d 1 a 592 was 8 $5.5 2 3. In 582.2. 2:. R 1 8 O O o 1 o 1 8 SE.— ally. 0 53585555 ozlvo Q 1 cm All/\ILOV IOd lNSOEBd 65 FIGURE 13. MECHANISM OF INHIBITION OF YEAST PHOSPHOGLUCOSE ISOMERASE BY HEXACYANOFERRATE(II). Phosphoglucose isomerase was assayed at a concentration of 22.8 ng/m1, in the presence of varying concentrations of fructose-6-P and hexacyanoferrate(II), as described in the Materiais and Methods section of this chapter. 66 um ' k L l L 1 l -5 0 5 1/ [FRUCTOSE-é- P] , mm" Figure 13 / 6rmm/ I L L I _[HEXACYANOFERRATE (I 0],! L C 10 67 substrate. The Km values are listed in Table l, as well as the Ki and Km/Ki ratios which we determined for PFK and PGI. The Km value of PGI for fructose-G-P (0.16 mfl) is 6.9 times greater than the Km of PFK for fructose-6-P (0.023 mfl); similarly, the K1 0f PGI for hexacyanoferrate (l.68 mg) is ll.2 times greter than the Ki 0f PFK (0.15 mfl). (The Km/Ki ratios for both enzymes were in the range of about 0.l.) TABLE 1. Km AND Ki VALUES FOR THE ENZYMES STUDIED.a Km FOR SUBSTRATE K1 FOR ENZYME Km/Ki FRUCTOSE-G-P GLUCOSE-6-P HCF(II) PIG-KIDNEY 0.023b --- 0.15C 0.15 PHOSPHOFRUCTOKINASE YEAST GLUCOSE-6-P --- 0.058-To.od DEHYDROGENASE YEAST PHOSPHOGLUCOSE 0.16e 0.8e 1.68f 0.095 ISOMERASE rKm and Ki are in millimolar units. b Massay, T.H., and Deal, w.c., Jr. (l975) Methods Enzymol. 42:99. C This work (2.0.5) d Noltman, E.A., and Kuby, S.A. (l963) in The Enzymes (Boyer, P.D., Lardy, H., and Myrback, K., eds.) 2nd Ed., vol. 7, p. 232, Academic Press, N.Y. e Noltman, E.A. (l972) in The Enzymes (Boyer, P.D., ed.) 3rd Ed., vol. 6, p. 271, Academic Press, N.Y. f This work (2.0.6) 68 Unlike PFK, PGI is not sensitive to sulfhydryl group oxidation so it can be assayed in a non-reducing environment. This allowed us to test the effect of hexacyanoferrate in the (III) oxidation state. As shown in Figure l2 the inhibition of PGI by hexacyanoferrate was identical in both reducing and non-reducing environments; this is further evidence that the inhibition does not involve oxidation. 7. Other Tests to Determine Whether Hexacyanoferrate(II) Binds to the Enzyme or Substrate. a. Analysis for Loss of Glucose-6—P Due to a Reaction With Hexacyanoferrate(II). Glucose-6-P was preincubated in standard assay buffer at a concen- tration of 0.5 mM, in the presence of l0 mM_hexacyanoferrate(II). At timed intervals after adding hexacyanoferrate, 100 and 200 pl aliquots were assayed for glucose-6-P as described in the Materials and Methods section of Chapter 3. The control experiment had no hexacyanoferrate— (II). The background absorbance at 340 nm of hexacyanoferrate(II) was subtracted from the total absorbance change, and the decrease in absorbance due to dilution of the assay by the glucose-6-P mixture was added. In these experiments there was a rapid initial l0% decrease in glucose-6-P and then no further change with time. This decrease was small, and may be due to technical problems created by the relatively high absorbance of the high concentration of hexacyanoferrate. Alternatively, it could represent a weak reverisble association of hexacyanoferrate with glucose-6-P at these relatively high concentra- tions of hexacyanoferrate. 69 b. Test for Ability of Fructose-6-P to Spare Glucose-6-P From Potential Binding to Hexacyanoferrate(II). The effect of hexacyanoferrate(II) on the activity of GGPD (using glucose-6-P as a substrate) was tested in the presence and absence of 4 mM fructose-6—P. The fructose-6-P is neither a substrate for G6PD nor does it affect G6PD activity. If hexacyanoferrate(II) were reversibly or irreversibly binding exclusively to the hexose phosphate substrate, then the excess fructose-6-P (4 mM) should largely react with the hexacyanoferrate(II) and leave a much larger fraction of the glucose- 6-P substrate unreacted. However, the added fructose-6-P was found to have no effect on the inhibition of G6PD. This suggests that the mechanism of the inhibition of GGPD, and presumably the other enzymes, by hexacyanoferrate does not involve either reversible or irreversible bindng of the substrate by hexacyanoferrate. 8. Preliminary Study Using Photo-Oxidized PFK PFK which had been photo-oxidized according to the method of Ahlfors and mansour (47) was shown to be still inhibited by hexacyano- ferrate(II). Since photo-oxidation inactivates the ATP regulatory Site, this is evidence that the regulatory Site does not have an essential role in the inhibition. E. DISCUSSION l. Mechanism of the Inhibition. The inhibition of PFK, PGI, and G6PD by hexacyanoferrate(II) appears to be unique since it does not involve an oxidation of either enzyme, substrate, or enzyme-substrate complex. Previous reports concerning the effect of hexacyanoferrates on other enzymes involved oxidation (39-44). The mechanism for the inhibition we observed seems to involve a competition with substrate by reversible binding of hexacyanoferrate(II) at, or near, the fructose-6-P (or glucose-6-P) binding site. 2. Molecular Basis for the Inhibition. This raises the question of how hexacyanoferrate(II) could compete with fructose-6-P and in particular, what structural similarity hexa- cyanoferrate(II) might have to fructose-G-P. The strucutral properties of hexacyanoferrate have been described in detail (49-50) and summa— rized (5l). The overall net charge of hexacyanoferrate is similar to that of the phosphate moiety of the hexose phosphate, but the molecular size of the hexacyanoferrate is clearly much greater than that of the phosphate moiety. This suggests that a substructure of the hexacyano- ferrate might be involved. The bond electron orbitals of hexacyanofer- rate are hybrids of the type, 3d24s4p3. The coordination positions of this complex with iron have octahedral symmetry, with all positions 70 7l equivalent (49). In contrast, the P-O bonds of phosphate have tetrahe- dral symmetry because sp3 hybrid orbitals are involved (49). The bonding of cyanide to iron is neither covalent nor ionic, but rather belongs to a bonding group called complex cyanides (49). The hexa- cyanoferrate complexes are remarkably stable (49). The evidence is strongly in favor of bonding through a metal-carbon bond rather than a metal-nitrogen bond (52), although the latter is not absolutely ruled out by present data (49). The C-N bond distance in cyanide ion is only l.05 A (49), which is considerably shorter than the P-O bonds in phosphate. In summary, the propeties of hexacyanoferrate do not Show any pronounced similarities to those of the phosphate ion. Their anionic character appears to be almost the only obvious common denominator structurally; but it does not provide an explanation for the structural specificity of hexacyanoferrate for these three enzymes. 3. Potential Active Site Directed Reagents. So far there have been no successful attempts to label the hexose phosphate catalytic sites of PFK, PGI and G6PD. Hexacyanoferrate(II) has a specific affinity for the hexose phosphate active sites of these three enzymes. It may be possible to use this pr0perty to label the active site of these enzymes by using the oxidized form of hexacyanoferrate. 72 4. Contrast with other Inhibitory Compounds, Especially Those Involving,0xidation. Hexacyanoferrate(III) has been shown (42, 44) to chemically modify the active sites of fructose-1,6-P2 aldolase, transaldolase, trans- ketolase, and ribulose bisphosphate carboxylase oxygenase. This process, which has been termed paracatalytic modification (42), involves a non-specific oxidation since the reaction is also carried out by a number of other oxidizing agents, such as tetranitromethane, H202, porphyrindin, and 2,6-dichlorophenolindophenol. Similarly, ferrate ion, Fe042', appears to be structurally similar to phosphate. It binds to the adenine nucleotide binding site of phos- phorylase b, producing a specific modification, through oxidation, of the active site (54). Ferrate ion has also been found to be a powerful specific inhibitor ofphosphatases (55), presumably also as a reuslt of binding to the phosphate binding site. All of these processes involve failry strong oxidizing agents, and they therefore differ fran the process described in the present paper. Since the experiments described in this paper were carried out in the presence of a strong reducing agent, the effects observed cannot be attributed to oxidation. Rather, they appear to be due to a specific, reversible binding of hexacyanoferrate(II) at, or near, the hexose- phosphate binding sites. CHAPTER 3 MEMBRANE-BOUND LIVER PHOSPHOFRUCTOKINASE AND METABOLITE-DEPENDENT REVERSIBLE SOLUBILIZATION 73 A. ABSTRACT Approximately 50 to 75% of the phosphofructokinase is fresh pig liver homogenates at pH 7.4 is particulate and associated with specific membrane fractions. The particulate phosphofructokinase is solubilized essentially immediately in fresh homogenates by addition of relatively high levels of fructose—6-P, fructose—l,6-P2, or glucose-6-P. Moreover, the solubilization is reversible; the soluble enzyme becomes particulate again as the added metabolites are converted to other products. Direct kinetic measurements of metabolite levels and part- iculate and soluble enzyme levels show a clear correlation between the disappearance of solubilizing metabolites and formation of particulate phosphofructokinase. The hexose monOphosphates mentioned above are rapidly interconvertible and quickly formed from fructose-l,6-P2 in the homogenates, so it was not possible to identify which of the indi— vidual hexose phosphates actually solubilize the particulate enzyme. AMP also rapidly solubilizes the particulate enzyme. The detegent Triton X-lOO solubilizes the particulate enzyme, but that process takes almost an hour, in contrast to the very rapid solubilization by the metabolites. In sedimentation equilibrium experiments with the part- iculate phosphofructokinase in sucrose solutions, the phosphofructo- kinase floats with a membrane fraction under conditions where nuclei and large protein aggregates sediment. The evidence thus far suggests that the enzyme binds to the plasma membranes. 74 75 Isoelectric focusing and DEAE-cellulose chromatography of the particulate and soluble pig liver phosphofructokinase suggest that both forms are identical in charge, and therefore probably do not differ in degree of phosphorylation. B. INTRODUCTION For a long time it was assumed that, with the exception of hexo- kinase in some tissues (58), virtually all the glycolytic enzymes existed solely in the soluble fraction of the cell. There is now quite a bit of evidence that certain glycolytic enzymes in contractile tissues may bind to contractile proteins; the list includes aldolase (59-6l), glyceraldehyde-3-P dehydrogenase (60), and phosphofructokinase (6l). This apparently explains observations of a number of years ago (62, 64) that phosphofructokinases in a number of different muscle (62, 64) and other contractile tissues (63) seemed to be particulate; they sedimented at very low centrifugal fields. In some cases (63) the proportion of enzyme sedimenting at low cen- trifugal fields increased if the tissue was allowed to age for several hours before the homogenization. Freezing also seems to cause or enhance particulate phosphofructokinase in several contractile tissues (62, 63), but not in rabbit liver (62). Kemp found particulate phos- phofructokinase in rabbit muscle but not in fresh muscle (62); in contrast, he found that freezing did not produce particulate phospho- fructokinase in rabbit liver (62). Two points just mentioned are espcially pertinent to the present studies. First, the phosphofructokinase in liver, a gluconeogenic tissue, is quite different from that in muscle, a highly glycolytic contractile tissue. Second, freezing and aging are treatments which 76 77 can sometimes give rise to the artifactual appearance of, or increases in, the amount of particulate phosphofructokinase (62). It appears that the particulate phosphofructokinase, aldolase, and glyceraldehyde—3-P dehydrogenase complexes in contractile tissues involve association of the enzymes with myofibrils, based on the known in vitrg_binding of these enzymes to actin (60, 65-67). This idea iS supported by the fact that the amount of enzyme bound to a particulate fraction is increased by electrical stimuli (59-6l). In addition to particulate phosphofructokinase formed by protein- protein associations, recent_in.vitrg studies showed that phosphofruc- tokinase and several other glycolytic enzymes bind to erythrocyte ghost membranes under certain conditions, namely low ionic strength and slightly acidic pH (68-77). The physiological significance of these results at first seemed questionable, because tight binding occured only at pH values of about 6.8 and below, and only at low ionic strength; also, at higher ionic strength (l50 mM KCl), all of the enzymes were dissociated from the membranes. However recent binding experiments with glyceraldehyde-3-P dehydrogenase suggest that it actually is membrane-bound in vivo, despite the in vitro results suggesting that it would be soluble (78). A particulate phosphofructokinase from yeast has recently been discovered (79), and has been found to bind to the plasma membranes. This particulate phosphofructokinase remains firmly attached to the membrane at high ionic strength and at high concentrations of sub- strates, and appears to be distinct from the soluble form of the enzyme. 78 In some preliminary studies of the effects of hormones on pig liver phosphofructokinase, we found that a large portion of the phospho- fructokinase activity in the homogenates sedimented at very low (900 x g) centrifugal fields. We then carried out an extensive study to characterize this phenomenon, using precautions to avoid possible artifacts. The results presented in this thesis show that a large portion of the phosphofructokinase in fresh pig liver homogenates is particulate, and appears to be membrane-bound. The results suggest that particulate phosphofructokinase may exist j__vivo in pig liver, and may be important in the regulation of the enzyme. .____‘ C. MATERIALS AND METHODS l. Reagents Most biochemicals were purchased from Sigma Chemical Co., except NADH, NADP, ATP, and other nucleotides, which were purchased from P-L Biochemicals, Inc. All chemicals were reagent grade or better. Imidazole was recrystallized twice in 3:10:5 (wzvzv) imidazole: chloroformzether. DEAE-cellulose, Triton X-lOO, ampholytes, and sucrose were purchased from Whatman, Rohm and Haas, LKB, and Mallinckrodt, respectively. 2. Tissues Pig liver was obtained from the Michigan State University Meats Laboratory. Most livers were obtained during the fall, winter, and early sring, from pigs which had been fasted l5 hr. Livers were removed from the animals within 20 min after slaughter, and immediately placed on ice. 3. Preparation of Liver Homogenates All procedures were done at O to 4°C. The homogenization procedure was begun promptly after obtaining the liver, usually within l5 min. Liver was cut into 2 to 4 mm pieces using a single-edge razor blade, and the pieces were washed in several changes of 0.25 M_Sucrose, to remove erythrocytes. The washed liver pieces were then transferred to 79 80 a 40 ml teflon-glass homogenizer (Potter-Elvehjem type) and 2 or 4 ml of homogenization buffer per gram tissue was added. For differential centrifugation experiments, 4 ml per gram was used; in other experi- ments, 2 ml per gram was used. The homogenization buffer consisted of: 50 mM imidazole-HCl, pH 7.4, 0.25 M_sucrose, and either 20 mM 2-mercaptoethanol or 5 mM dithiothreitol. Liver was homogenized with 5 strokes of a size-C pestle, rotating at about 900 RPM. The homogenate was then filtered by pouring through 400-mesh nylon cloth, or several layers of cotton gauze. 4. Differential Centrifugation This procedure was adapted from the method described by Fleisher and Kervina (80) for isolating subcellular organelles from rat liver. Figure l4 outlines the general procedure, rotors, and centrifugation conditions, and volumes of buffer used to resuspend pellet fractions. The homogenates were prepared from l2 to 48 grams of liver, using 4 ml of buffer per gram, as described previously. The standard buffer used for resuspending pellet fractions contained the same components as the homogenization buffer, plus the addition of l_M MgCl2. This buffer was used in all procedures after homogenization, unless otherwise noted. Centrifuge tubes for the Sorvall SS—34 rotor were the 29 x l02 mm polycarbonate types; tubes for the Beckman 42.l rotor were the l x 3.5” cellulose or polyallomer types. For resuspending the lOOO x g pellet (and other fractions enriched in plasma membranes), a Dounce homogenizer was used, with a loose-clearance pestle (80). For most other pellet fractions, teflon- glass homogenizers were used. 81 FIGURE I4. PROCEDURE FOR DIFFERENTIAL CENTRIFUGATION OF LIVER HOMOGENATES. The procedure was carried out at 0 to 4°. Other experimental details are described in this Materials and methods section (3.6.4). 82 FILTERED HOMOGENATE (33. 50 m1) 1000 x g SS-Bh rotor 10' 5250 RPM \/ PELLET SUPERNATANT Resuspend in 20 ml buffer with Dounce . 25,000 x g ; SS-jh rotor homogenizer. 10. 16,500 RPM V/ PELLET SUPERNATANT ResuSpend in 10 ml buffer with Teflon/ 120,000 x g 42.1 rotor glass homogenizer, 70' 40,000 RPM // PELLET SUPERNATANT Resuspend same as previous pellet. Figure 14 83 5. Sucrose Step-Gradient Centrifugation of Particulate Pig Liver PFK. Suspensions of l00 x g pellet fractions fron liver homogenates were prepared as described previously, and mixed with concentrated (2.4 M) sucrose (in the standard buffer), to bring the final sucrose concentration to l.46 or l.6 M_(54.8 or 50% [w/v], respectively). Aliquots (about 30 ml) of this suspension were then transferred to l x 3.5" cellulose nitrate centrifuge tubes. The tubes were then filled carefully with buffer and centrifuged in the Beckman SW-27 rotor at 25,000 RPM for 1 hr (at 4°). Figure 24 shows the appearance of a tube before and after centrifugation. Gradients were then fractionated as follows: The clear buffer on top was removed with a pipet and discarded. The floating band, at the heavy sucrose interface, was collected with a stainless steel scoopula. These bands were pooled and resuspended in 2 ml of the standard buffer per initial gram liver, in a Dounce homogenizer. The liquid below the floating band was poured off, and the pellets were combined and resus— pended in a teflon-glass homogenizer, using about l ml buffer per initial gram of liver. The buffer used to resuspend this pellet frac- tion was modified to contain 3 mM MgCl2. in order to stabilize the nuclei (80, 8l), which were in this fraction. 6. Enzyme Assays. Phosphofructokinase activity was determined according to the stand- ard procedure of Massay and Deal (82), with the modifications that imidazole-HCl, pH 7.4, was substituted for TRIS buffer, and the volume of the reaction mixtures was adjusted with H20 to allow l0 to 50 u] aliquots of enzyme solution to be added to the cuvettes. The 84 spectrophotometer slit-width was set a maximum (2 mm) to minimize background interference when assaying turbid fractions. The relatively high concentrations of fructose-6-P and ammonium sulfate in the cuvettes (4 and 40 mM, respectively) ensured that the added phosphofructokinase was solubilized and maximally activated. Lactate dehydrogenase activity was determined according to the method of Kornberg (83). This enzyme, which is higly active and pre- dominantly soluble in liver, was assayed as a control for unbroken cells and soluble supernatant protein carried over in pellet fractions. Fructose-l,6-diphosphatase activity was determined spectrophoto- metrically according to the method of Racker (84). 7. Determination of Relative Particulate and Soluble Enzyme Activity Distributions in Pig Liver. Aliquots (l to l.5 ml) of homogenates, prepared as described pre- viously, were incubated under various conditions, transferred to l2 ml centrifuge tubes, and centrifuged in the SS-34 rotor at 8000 RPM (l2,800 x g at for 5 minutes, at 0 to 4° or 23°. Superna- rmax) tants (soluble fractions) were carefully transferred to calibrated tubes. Pellets (particulate fractions) were resuspended in l to l.5 ml of standard buffer (3.C.3), with a Teflon pestle. Aliquots (l0 to 20 pl) of the fractions were then assayed for PFK and LDH activity as described previously. 85 8. Metabolite Determinations a. Reagents perchloric acid (HCl04), 6 N, (stock 70% [w/w] = ll6.9 N) K2CO3, 2.5 M triethanolamine ° HCl, pH 7.6/MgClz, 400/l0 mM triethanolamine ° HCl, pH 7.6/EDTA, 400/40 mM NADH, 20 mM, in l mM_Na0H NADP, 50 mM ATP, ADP, AMP, l00 mm, pH 7 glucose, l00 mM phosphoenol pyruvate (PEP), 40 mM sodium pyruvate, 80 mM F-6-P, F-l,6—P2, TOO MM Enzymes used as reagents are listed in Table 2. Commercial stock enzymes were diluted with 3.2 M ammonium sulfate and H20, as shown in Table 2, and stored in screw-capped vials, at 4°. The triethanolamine buffers (with MgClz 0r EDTA) were described by previous workers (85, 86), and were also stored in amber bottles at 4°. b. Preparation of Perchloric Acid Extracts of Liver Homogenates. Perchloric acid has been commonly used by previous workers (85,86) in preparing tissue extracts for metabolite determinations. The pro- cedure was done at 0°. Aliquots (l to l.5 ml) of homogenate, previous— ly incubated with various metabolites, were diluted into 2 volumes of cold (0°) 6 fl_HClO4, in l2 ml glass centrifuge tubes, and immediately mixed, using a vortex mixer. (The final HClO4 concentration was 4 N,) Tubes were then centrifuged in the 53-34 rotor at 8000 RPM for 10 86 web arcane OP .mpmw_:m EzwcoEEm .m<~ .mpnmp mcp eo wwwm psmwe asp co umumuwvcw mm .om: new wpmw_:m EswcoEEm z.m.m new; uwpspwu mew mxoopm PmqumEEoo msp .mco_p:8vecmpmu mpwpoampma Loy m20wumegcmocoo Pmcww vmg_mwu .vmgmwp mew mmex~zw xoopm meoLmEEou we mmwuv>_¢om uwwwomam use mcowpmepcwocoo msg— mFomse - com co, cap ©_.o om__ N.m m._ awe amm=_¥oaz mFUmze mmm mme om mme m.o cum _.N o.o_ pwnnma amneamoee»;ao abepoao wFumze com 0mm om mmm m.o woe N.N o.o_ Began; ame:_¥ apa>acza om mew mromze mmwcmmoguzcmo ¢1Fogmoxpwua - owe 0mm \oem N.N \Nmmm N.m m.o_ BLQBML \wmmeaEOmH mpmgamoeaamOmFLe mFomze wwaoup< - com oo_ N_ o._ N_ N.m o.o_ ownnme apagamogamen-m._-aWOBU=LL oe_ o4“ oo_ com em.o 0mm m.~ ¢©.m ammo» wmmaasomH amoo=_mogamo;a - com ow me_ mm.o ome N.m m.m “mama mmw=_xoxaz - 0mm om_ om_ m.o com N.m om._ ammo» mmacmmoeexgao a-m-amoo=_w .qm: 1m< meaNcm _E\: Pe\me Ame\:q TNEV A_E\mEo z N.m Bacon zoee_pow .ocou .ocoo wumzom mz>sz masz.e =_ uwza._ _; Ssz no monHDJHo oz< monHmL 1vwom merrammme xa uwcw::mpmu mm; ma1o.mnomoposee .mH.u.m co_uomm cw umnvuomwu vogpme eomm:\_o:oga mcp mcwm: umcwELmumu mam; meowpmepcmocoo mumruxcongmo m>wpm_mm .Awmm _muoe .oomHV Lopow_—oo coruumew creme uopam cm mcvm: umuom_Foo mum: Amaoru1o~V meowuomeu .:WE\_E mm.o Lo ape; zoPe m an .pcmwumem Fonz.m ¢.o op _.o mecPF _E omm w 59?: emu:_m we: :E:_oo mg» .Am.m_.o.mv mnosumz 5cm mpmwgmumz cw nonweomwu mm ._1xwzoo co umgamcmopmeocgo we: waspxwe asp .mmmaoa cpwz wrapx_e cowpomwr N¢1m.m1mmopoare esp mcwpmwep mew< .onH1aogo ooopromoo mo oocpeLopoo moo; No1m._1omopoorp oco No1m.m 1omopooeu .mooeo mm op .oomoorooo mo: mas—o> ooppootp osp poooxo .mp opompm op ocmmop mcp :p ooopromoo he; ocam mop oogooemopoEoegooL mum: ass—co p1xo3oo pmppp mop Soup AQM1Nm .moov mcoppooum oopooo asp .>Iomno ooopeomoo mo .moppp>ppoo Mme opozpom oco opopooppeoo mopcpELopoo Low oom: oco .m x ooo.o_ po oomop 1pepcoo .cmxop om—o mew; mpoooppo Fe moo .Aw.u.mv mcoppoopELopoo mpppooopoe to» oom: oco .Loxmpxo> o co ooxps .oopu:.a o opoo po _5 p :p oopoppo .coxop mew: mpoooppo FE m.o .googm ocp :o exogm .m_o>rop:_ ooepp ocp p< .23 N po coppoepcoocoo o op o1o1omopoogp mcpooo xo pcoepeooxo ocp mcpccpmoo onopoo 0mm po ope m Lop opmooocp op oozoppo ooo moogpoz ooo mpopgopoz :p ooopeomoo mo Lo>pp mpg Soap oouoooeo we: opocomoEo; < .zophmpomo a1m1mmohozmm mmpm< xmo mhm omzoppou .m1o1mmOHuamm_EE N >m Axmov mmzoo .om mmzwpm ll7 om mtzopu 5.6822. op a: 2... K 2:22 55 25 E: R: 8 8 d 1 E 1 co ‘ cm 8 R d H... d d 0.. 00 an» 2.3: :39: :5: 28328: - 55‘}. m.“ QN m.~ 9m “nouvzunaonoo aauomnn amnaov WU] ll8 7. Specific Solubilization of Particulate Phosphofructokinase by Hexose-Phosphates and Nucleotides. Addition of a number of hexose-phosphates and certain nucleotides to homogenates was found to affect the relative distribution of the soluble and particulate forms of phosphofructokinase. Detailed studies were carried out with fructose-6-P (Figure 2l), fructose-1,6-P2 (not shown), and AMP (Figure 22). Solubilization constants (defined in the legend to Figure 2l), were determined for these and other metabolites, and summarized in Table 5. The experiments shown in Figures 2l and 22 were repeated several times with livers obtained during the fall and winter, and no significant differences were found in the solubilization constants. However the total phosphofructokinase activity in different livers varied as much as l.8-fold. We are not certain what causes the variation in activity. a. Solubilization by Fructose-6-P. In a homogenate with no added fructose-6-P, 52% of the phospho- fructokinase activity sedimented with the l0,000 x g pellet (Figure 2l). Upon addition of fructose-6—P to the homogenate aliquots, the fraction of phosphofructokinase activity in the supernatants increased, reaching a maximum of 80% at about 200 pM fructose-6-P. Part of the unsolubilized 20% activity in those pellets could have been due to included supernatant. This residual pellet activity varied from l0 to 25% generally. The solubilization curve for fructose-6-P has a general hyperbolic shape. Glucose-6-P and fructose-l,6-P2 gave results very similar to those for fructose-6-P. Since those compounds are interconverted ll9 .mpoooo coppoNpppoopom Eoepxoe ocp po ppm; memo; coppoppcooooo ocp mp m.oo ozp .coppowpppoopom Eoepxoe Apopoepxopooo mcp>pm o1©1omopoopp po coppoppooocoo pmogop ogp mp .o._u .coppmppoooooo coppoNpppoopom poepxoe on» .coppoopp comm :p Amppo: xgopppopo opv app>ppoo popop osp m3ogm :oopm one .ooopELopoo opp; moppp>ppoo omocpxop Ioopposomoco opoopom oco opopoopppoo oco .m x ooo.o_ po oomoppppcoo .ooxoppo> apxopoo mooop ogp .ooooo coop one; o101omopoopp po mpoooEo mcprm> .Aommv opopopooEop Eoop :ooop op .mopocpe e poooo ocopm op oozoppo ooo mooop omoppepcoo FE Np op ooooo moo; opocomoEo; Lo>pp mpo ompooopo apsmopp po mpoooppo FE woo .o1m1mmOHuzmu >m mmTF pmc spwz mcowpwucoo gmpwarm cmucs vase» mmogp mew .mwmmgpcwcma cw umpmw_ .mcowpumcm asp we mpcmcanoo Lm_:p_mon:m pcwcweovmca mgh .cowpm>wpom cu mzv Lo cowpwnwgc? mo Fm>oewg op mac on e_zoo mvcp gmcpmgz mczm poc «cm m3 .pr>wpow mcwpgmpm asp :mgp qummcm awfi mm: pmppma use econ mcwpmo—m mcp Eocm umcm>oomc pr>wpom _mpou age .abw>wpom mmmcwxouozcmoggmosg o: um:_mpcoo ucmn mcwpmo_$ mgp soan ecu w>onm e_zcv_ one .co_umm=wwcycmu cmumm can mgomma pcm_umcmnampm a mo mocmcmmaam asp mzogm Ewcmm_u mch .Am.o.mv mnogpmz new m_m_cmpmz cw uwn_commu mw mczumooca _mp:mswcmaxm mch .Pmsswa m x ooo_ Lo onHm omzH>mmn .moawunsoe «Emufiav Qz33“? E .5 SE a moofi W mmoxosm z m.~ magm «manna zH yuagmm w x ooo.~ amazmmmsmmm mama—m 2H mmomoam 2 $6 134 These results show conclusively that: (l) the particulate phospho- fructokinase is not an extremely high molecular weight particle; there- fore its presence in membrane fractions is not a result of being highly self-associated; (2) that the major portion of the particulate enzyme is not bound to nuclei; and (3) that the particulate phosphofructo- kinase is membrane-bound. The basis for conclusions one and three is that the density of protein is about 1.33 g/cm3, so the enzyme must be membrane-bound to float as a particle with a density in the range of l.l8 g/cm3. l2. Attempts to Reconstitute the Phosphofructokinase-Membrane System. The floating bands fran the sucrose step-gradient experiments were found to contain the major portion of the particulate phosphofructo- kinase (Figure 24). The major subcellular components in this fraction are nominally the plasma membranes and heavy mitochondria (80), so it seems likely that the phosphofructokinase could be bound to one or both of these fractions. However when the individual membrane fractions are isolated, neither of them contains phosphofructokinase activity. This is most likely due to a gradual dissociation of the enzyme during the repeated centrifugations in the presence of EDTA, which is part of the isolation procedure. As one approach to determine the membrane to which phosphofructo- kinase binds, we carried out incubation experiments with mixtures of purified or partially purified pig liver phosphofructokinase and various subcellular membrane fractions, as described in Materials and Methods (3.C.11). Most experiments were done with 0.5 ml incubation 135 aliquots, containing 0.15 units/m1 phosphofructokinase, and 5 to 10 mg/ml membrane protein. In a typical experiment, the control (without membrane) showed 44% of the phosphofructokinase activity in the 27,000 x g pellet. In the mixture containing purified plasma membranes, 64% of the activity sedimented at 27,000 x 9. However the nuclei, heavy mitochondria, and major mitochondrial fractions also showed a substantial amount of sedimenting activity, (55, 52, and 54%, respectively). Thus a large amount of non-specific precipitation occured in the reconstituted systems, and the amount of sedimenting activity in any fraction was not significantly above the background levels. We also observed no enhancement of binding in experiments with 120,000 x g supernatant included in the mixtures, or in experiments with crude pellet and supernatant fractions. These results suggest that perhaps another cellular component (5), in addition to enzyme and membrane, is required for binding to occur. The other component might be inactivated during the fractionation procedure. Or perhaps the membrane receptors for phosphofructokinase become solubilized during the fractionation. The membrane would then be unable to bind phosphofructokinase, or the solubilized receptors might compete with membrane-bound receptors for the binding site on the enzyme. 13. Analysis for Different Phosphorylated Forms of Phosphofructokinase by DEAE-Cellulose Chromatography. Previous workers have found that phosphofructokinases from a number of different species and tissues can undergo in vitro phosphorylation 136 .mpcwewcmaxm wpmgmamm ozp asp Lo» mw~_$oga covpz_m mzp mzogm :amgm mce .5 umNPFPns—om mschm muszowuqu wgp mm mm: mamm esp wosamgmoumEoLgo cwsp ucm .mcwmaoga mpnz_om gmgpo w>o=mL op .vma wmoP=FFmoumw_ mwa mpn:_om .Am.o.mv mcogpoz use mpmwgwpmz cw umn_commu mm :Ezfioo mmo_:~_moumv_ awn 502+ umpumcpxw mm; mmmcwxouoag$osgmosa mum—:uwugmm .m QMNHAHmDAQm .mmH4 wHa mHIa :0 L030_ 0» 0x000 zp_>0000 0:0 :_ 000:0 0:0 000000 0>0s 005 0:0 .0000000000000 0:0 :0 000000000? 0>0g 005 £000; .—0wg0u0e 00p0_:m000 00 000000 0mL0_ 0 003 000:0 .Au 000 m 00csmwmv 0u0zp0 0:0 00 00E:_0> L0mL0P 0000000000 00:0E_000x0 050 :H .030000 050 An 00p0o_00_ 0L0 000000000 xuw>wpo0 £000 0;» 00 AHQV 0:000 000000P0000 0:0 .x_0>00000000 .0000F0 000 000_:__0oum0000 0gp .A_.u.mv »_0:00>000 000000000 00 .00000PF00 0003 0:0_00000 _E m_.o 00;» 0:0 .0; OF 000 0000000000F0 00 0020__0 0002 0:50—00 00H .Ao_.u.mv 00px—05050 0-0 :0 :00; .PE N.o 00 050F0> 0 :0 .omm p0 0000 003 00000000 e_Luo0F000H .A0.m.u.mv 000 000_:__001mH4 me no wszzoom onHQMAmomH .om mmszm 141 All/\llOV )Hd BAIIV'IEIEI ) S 8 T 00 003000 20.85; .0. 0.5.0 fl 3 3 ”V3 0 \ O 1 N\ 8.0 u a I 4 0. ..... 1 .o ..... o... ...O. I a 1 0.. ...c we 0.0 0.0 wé 9m ~.m ed :0 E. DISCUSSION 1. Experimental Effects and Correlation with In Vivo Metabolite LEI/Elfi- A new property of pig liver phosphofructokinase has been discovered and characterized. The results show that more than half of the phos- phofructokinase in fresh pig liver homogenates is particulate, and that the level of particulate phosphofructokinase is highly sensitive to the levels of certain hexose and nucleotide metabolites. The results are consistent with the formation of particulate phosphofructokinase when the levels of hexose phosphates are low, a condition where the glycoly- tic rate should be decreased; the results are also consistent with the formation of soluble phosphofructokinase when the hexose phosphate lev- els are high, condition where the glycolytic rate should be increased. Measurements of metabolite levels and of particulate phosphofructo- kinase as a function of time show a direct correlation between particu- late enzyme and decreased levels of glucose-6-P and fructose-6-P. Flotation experiments provided evidence that the particulate phospho- fructokinase is membrane bound, mose likely to the plasma membranes. One key question is whether the results observed could occur at metabolite concentrations which exist i vivo. Although data are not available on the levels of metabolites in pig liver, some indication of order of magnitude can be obtained from results with rat liver, from fasted and fed rats (Table 4). Of special interest is the fact that 142 143 the_ifl_vivg levels of the hexose-phosphates are higher in fed animals than in fasted animals. Furthermore, fructose-6-P and glucose-6-P vary the most, showing a two-fold increase in fed liver over that in fasted liver. If similar hexose phosphate levels were to occur in pig liver, the fructose-6-P in fed animals would solubilize approximatley 60% of the particulate phosphofructokinase (estimated from Figure 18), while in fasted animals, approximatley 30% of the enzyme would be solubilized. The glucose-6-P concentration in fed animals would solubilize all the phosphofructokinase, while that in starved animals would solubilize somewhat more than half of the particulate enzyme. However, glucose- 6-P is not an allosteric activator of the enzyme and it may not be a true solubilizer; rather, its apparent effect may be due to fructose- 6-P produced fron the glucose-6-P. 2. The Membrane to which Phosphofructokinase Binds. The sucrose gradient studies show that phosphofructokinase is associted with a membrane, and that its presence in the particulate fraction is not a result of extreme self-assocition, or association with another high molecular weight macromolecule which would sediment with the particulate fraction. The simplest explanation of the results is that phosphofructokinase binds to the plasma membranes. This requires an explanation for the presence of phosphofructokinase activ- ity in both the nuclear and microsomal fractions. Previous workers have shown (92, 93) that plasma membrane-bound enzymes fron liver (92) and other tissues (93), show a bimodal distribution of activity between the nuclear and microsomal fractions. Presumably the plasma membranes 144 must form either very large or very small fragments when the liver is homogenized in 0.25 M_sucrose; thus the plasma membrane fragments sediment with nuclei and microsomes, but not with intermediate sized fractions such as mitochondria and lysosomes. Although some mitochondria also sediment with the nuclear fraction, it seems unlikely that liver phosphofructokinase binds to mitochondrial membranes, because we observed no phosphofructokinase activity in the major mitochondrial fraction obtained by differential centrifugation. The flotation experiments showed that the major portion of the phospho— fructokinase does not bind to the nuclear membrane, because most (80%) of the phosphofructokinase activity floats, while only about 20% is found in the nuclear pellet. 3. Why the Membrane-Binding Could be Significant. This study documents the phenomenon of liver phosphofructokinase association with the plasma membrane, and shows a clear correlation of increased binding with decreased hexose—phosphate metabolite levels. Now that the phenomenon has been characterized, further studies can be done to determine what possible significance these properties might have in the regulation of phosphofructokinase_1fl_yiyg. Binding to the membrane certainly might affect the activity of phosphofructokinase, or it might affect the pr0peties of the membrane. During conditions when hexose phosphate metabolites are elevated, and phosphofructokinase activity would need to be increased, the enzyme is mostly in the soluble form, where it could diffuse freely in the cytoplasm and be more acccessible to its substrates. When hexose phosphate levels are decreased, and the phosphofructokinase activity would need to be 145 decreased, the enzyme is mostly in the membrane-bound form, and perhaps less accessible to the substrates. This would be especially pronounced if the binding occured at or near the active site of the enzyme, where it could prevent the binding of substrates by a competitive mechanism. CHAPTER 4 STUDIES ON PIG HEPATOCYTES 146 A. ABSTRACT Pig hepatocytes were isolated in order to study the interaction of liver phosphofructokinase with membranes. Hepatocytes were isolated by two methods: treatment of liver slices with collagenase and dispase (104), and a modified collagenase perfusion technique. The tissue slice incubation procedure gave hepatocytes with a high degree of intactness and viability (75 to 90% by trypan blue exclusion); however the yields of viable cells were low (less than 105 cells/g liver). The collagenase perfusion technique gave much greater yields (2-5 x 106 cells/g liver), although the percent Viabilities were more variable, ranging from 30 to 80%. The phosphofructokinase activity in the isolated hepatocytes was found to be much lower than that in whole pig liver, and it appears that most of the enzyme activity is lost during the hepatocyte isola- tion procedure. What activity remained was found in the soluble frac- tion. These results differ markedly from those observed with whole pig liver, where the total units of homogente phosphofructokinase activity is 5 to 15 times greater, and where 50 to 75% sediments at low (1000 x g) centrifugal fields. In incubation experiments with isolated pig hepatocytes (37° for 20 min), the total phosphofructokinase activity increased 23% in the pres- ence of 40 mM glucose, and decreased 15% in the presence of 20 mm lactate. Since phosphofructokinase activity was measured at high 147 148 fructose-6—P concentration, the changes in activity must be due to changes in the maximum velocity or in the concentration of active enzyme. B. INTRODUCTION In the studies previously described in Chapter 3, we found that pig liver phosphofructokinase in whole liver homogenates exhibited a unique, reversible, metabolite-dependent, membrane binding phenomenon. We observed that variations in the concentration of the substrate fructose-6-P, within physiological levels, could regulate the degree of phosphofructokinase membrane binding. This raised the possibility that the membrane binding phenomenon might occur_i__viyg. Our next interest was to attempt to answer the questions of whether phosphofructokinase might actually bind reversibly to membranes 1_ vivo, and the possible physiological significance. Isolated hepatocytes seemed to be the best system to study such a possibility. We first considered studying the phosphofructokinase membrane binding phenomenon in hepatocytes from much more convenient sources, such as rabbit or rat liver. However, in studies on whole liver from these animals (described in Chapter 5), we found their liver phospho— fructokinases were quite different, and did not display the membrane binding properties we observed with the pig liver enzyme. It was therefore necessary to use isolated pig hepatocyes to study the phos- phofructokinase membrane binding. Isolating hepatocytes from large animals has several difficulties not encountered with smaller animals. Perfusing the entire liver by the portal vein is usually not practical, and therefore perfusion must 149 1'1 150 be done with a lobe, or a peripheral portion of a lobe, which is usually not as efficient. Also, the large livers have a much greater amount of collagen, making it more difficult to disperse the cells. In spite of the difficulties however, collagenase perfusion techniques have been used successfully, by Clark_et._l. (102, 103), for isolating hepatocytes from lamb (102), and adult minature pigs (103). Miyazaki _gt_gl. (104) reported the isolation of human hepatocytes from liver biopsies, by enzymatic treatment of liver slices with collagenase and dispase. They found that dispase (a neutral protease isolated from Bacillus polymxa 105), although not effective by itself on liver, appearaed to aid the liver dispersal when used subsequently to the collagenase treatment. For our studies we used hepatocytes isolated by collagenase and dispase treatment of liver slices, and by collagenase perfusion of a peripheral portion of a liver lobe. We then investigated the effects of glycolytic and gluconeogenic incubation conditions on the level and distribution of the hepatocyte phosphofructokinase. C. MATERIALS AND METHODS l. Tissues and Reagents. Most chemicals were obtained from the sources described in the Materials and Methods section of Chapter 3, and were reagent grade. Collagenase (type CLS II) was obtained from Worthington. Porcine insulin (crystalline, 24 I.U./mg), glucagon (crystalline), and dispase (neutral protease, type IX) were obtained from Sigma Chemical Co. Dispase was assayed according to the procedure described by Kunitz (106), with casein as the substrate, and in Hanks buffer plus ZOTMM HEPES, pH 7.4. (We found that the presence of Ca2+ and MgZ+ ions was essential for Optimum activity of the enzyme.) A unit of protease activity (107), is the amount causing an increase in absorb- ance at 280 nm of 0.001 absorbance units per minute. Pig liver was obtained from the Michigan State University Meats Laboratory, fran animals which were starved 12-15 hr before slaughter. About 40 g of a peripheral portion of a lobe was excised (102) within 20 minutes after sacrifice of the animal, and placed on ice for trans- port to the laboratory. When the collagenase perfusion technique was used, the lobe was immediately perfused with about 200 m1 of Ca2+ -free perfusion buffer (108), containng 0.l mM_EGTA, at 0°, according to Forsell and Schull (109). 151 152 2. Isolation of Pig Hepatocytes. a. Enzymatic Treatment of Liver Slices. The basic procedure described by Miyazaki gt_al. (104) was carried out, with the following modifications. Liver slices were incubated in 6 g aliquots, in 500 ml Erlenmeyer flasks, with 20 ml of buffer. Krebs-Henseleit buffer (110) was used for the collagenase and dispase treatments for most experiments. In several experiments, Dulbecco's modified Eagles medium (serum-free) was used. The flasks were incubated at 37° on a shaker-water bath, shaking at a rate of about 60 cycles/min. The flasks were gassed at 10 min intervals with 5% 002/95% 02, to maintain the pH at 7.4-7.5. For most experiments, the incubation time was 20 min. b. Collagenase Perfusion. This procedure was carried out in Dr. Lee Shull's laboratory at Michigan State University, with the help of Mr. James Forsell, who collaborated with us on this section of the project. The method has been previously used for the isolation of bovine hepatocytes (109). This techniques uses the basic perfusion apparatus, buffers, and general technique described by Seglen (108) for isolation of rat hepatocytes, with modifications for perfusing a peripheral end of a lobe of a pig liver. The general procedure is as follows. (1) Pig liver was obtained from the Michigan State University meats laboratory, fran animals which were fasted 15 hr before slaughter. As soon as the liver was removed from the animal, a peripheral end of a lobe (about 60 g) was excised by a straight transverse cut, using a scalpel with a no. 11 blade. The liver specimen was then immediately 153 perfused with 250-400 ml of ice-cold Ca2+ and Mg2+ -free saline (108) plus 0.5 mM EGTA. Perfusion was done with a 50 m1 syringe, with a plastic tapered tip, fitted into a major blood vessel. This perfusion was continued until the blood was removed, and the liver had a tan color. (2) The liver specimen was then transported to the laboratory (in the same buffer, at 0°), and trimmed to a weight of about 40 g, by another transverse cut, parallel to the first cut. Subsequent perfusions were done on an apparatus described by Seglen (108), using plastic canula fitted into major blood vessels. During the perfusions, the canula may be moved to other blood vessels, to ensure that the perfusion buffer reaches all parts of the liver. The liver was first perfused at 23°, with 200 ml of Ca2+ and Mg2+ -free perfusion buffer (108) (without EGTA). (3) The liver was then perfused at 37°, with 125 m1 of collagenase buffer (108), containing 0.6 mg/ml of crude collagenase (Worthington, CLS II). The buffer was equilibrated with 95% 02/5% C02 gas mixture, using a sintered glass aerator. Perfusion was done in a recirculating mode (108) for 50 to 60 min. (4) After collagenase perfusion, the liver was dispersed into a cell suspension, using the combing technique described by Seglen (108). (Generally, the more efficient the perfusion, the greater proportion of the liver could be made to go into suspension.) Parenchymal cells were then isolated as follows (108). The crude cell suspension was centri- fuged in a clincial centrifuge with a swinging bucket rotor, at 50 x g for 3 min. The pellet was gently resuspended in about 50 m1 of Krebs-Henseleit buffer (110) at 0°, and centrifuged as before. The 154 resuspension and centrifugation steps were repeated at least 2 more times, until the supernatant was clear. Percent viability was deter- mined by trypan blue exclusion (108). The isolated parenchymal cells were then suspended in KHB plus 2% bovine fraction V albumin (Sigma), and kept at 0° until beginning experiments. Most experiments were begun promptly after the hepatocytes were isolated, however the cells could be stored at 0° for at least 24 hr without losing viability. 3. Incubation of Hepatocytes. Incubation experiments were done with 4 or 5 ml aliquots of hepato- cyte suspension, in 25 ml Erlenmeyer flasks. hepatocytes were suspend- ed at a concentration of about 5 x 105, in Krebs-Henseleit buffer (110) containing 1.5 NM CaClz, 0.8 mM.MgSO4, and l-2.5% bovine fraction V albumin. Flasks were kept at 0° until beginning the incuba- tion at 37°. After adding various substrates and hormones to be tested, the flasks were transferred to a shaker-water bath, and incu- bated at 37°, shaking at a rate of about 60 cycles/min. The flasks were gassed with 5% 002/95% 02 for l min while incubating, and stoppered. The flasks were gassed again after the first 10 min of incubation. For most experiments, the incubation time was 20 min. After incubating at 37°, the flasks were placed on ice, and the homogenization and enzyme assay procedures were started promptly. 4. Determination of Phosphofructokinase and Lactate Dehydrogenase Activities in Hepatocytes. All procedures were done at 0 to 4°. Hepatocytes suspensions, incubated as described previously, were transferred to 12 ml centrifuge 155 tubes (Sorvall) and centrifuged in the Sorvall SS-34 rotor at 2500 RPM for 3 min. Cell pellets were then suspended in 0.5 ml of homogeniza- tion buffer, containing 50 mM,imidazole-HCl, pH 7.4, 0.25 M_sucrose, 1 mM M9012, and 5 mM DTT. The suspension was transferred to a 2 ml Teflon-glass homogenizer and homogenized with about 10 strokes of the pestle. Homogenates were then centrifuged in the SS-34 rotor at 8000 RPM (10,000 x g) for 5 min. Supernatants were transferred to separate tubes, and pellets were resuspended in 0.5 ml of homogenization buffer, using a Teflon pestle. Both supernatant and pellet fractions were assayed for phosphofructokinase and lactate dehydrogenase activity as described in the Materials and Methods section of Chapter 3. D. RESULTS 1. Yields and Viabilities of Pig,Hepatocytes Isolated by Enzymatic Treatment of Liver Slices and By_Collagenase Perfusion. Hepatocytes were initially isolated by incubating liver slices in the presence of collagenase and dispase, according to the basic pro- cedure of Miyazaki gt al. (104), as described in Materials and Methods. The optimum concentration of dispase was about 8 mg/ml. Yields of hepatocytes from this method were low, in the range of 105 cells/g liver. This was markedly lower than the 2 x 106 cell/g yields reported by Miyazaki 2E.§l- (104) for human hepatocytes. The percent Viabilities were usually good, in the range of 75 to 90% (by trypan blue exclusion). In two preparations, we used Dulbecco's modified Eagles medium as the isolation buffer. In both preparations the via- bilities were 90%, however the yields were not improved. In order to obtain greater amounts of hepatocytes needed for incubation experiments, we next tried the collagenase perfusion technique, as described in Materials and Methods (4.0.2.b). This method gave much greater yields of parenchymal cells, averaging about 2.5 x 106 cells/g liver, and a 40 g lobe section yielded about 108 cells. The percent Viabilities were more variable however, ranging from 30 to 80%. Repeating the washing and centrifugation steps resulted in higher percent Viabilities, at the expense of total yield. 156 157 The yields were enhanced when the total anount of collagenase used was increased from 125 to 200 mg. 2. Level and Distribution of Phosphofructokinase and Lactate Dehydrggenase in Isolated Pig,Hepatocytes versus Whole Liver. Once we obtained sufficient quantities of liver parenchymal cells, we were next interested in determining the level and subcellular distribution of phosphofructokinase. As a control, lactate dehydrogenase, a soluble enzyme, was also assayed. Homogenization was done with a Teflon-glass type homogenizer, and in the presence of isotonic (0.25 M) sucrose, as described in the Materials and Methods section of this chapter. Light microscopic examination of homogenate fractions showed there were no unbroken cells. Homogenates were prepared with 2 x 107 cells and 0.5 ml buffer. In four different hepatocyte preparations, the average homogenate phosphofurctokinase activity was 0.0165 U/ml. This was about l/15 of the average value of 0.25 U/ml for whole liver homogenates (l g tissue/2 ml buffer). The lactate dehydrogenase activity was also lower in hepatocyte homogenates, by about 8-fold. These large differences in activity are much greater than the 1.5 to 2-fold differences predicted by the difference in tissue concentrations. This suggests a large amount of the enzyme activity must be lost during the hepatocyte isolation procedure. Previous workers have also observed losses in constitutive enzyme activities in hepatocytes (108, 102, 112, 113). In some cases (108, 102) the losses were attributed to leakage of protein through the plasma membranes. (Presumably the membranes may have been 158 slightly damaged.) Other workers (108, 113) have found evidence of protein degradation in hepatocytes, and perhaps this may be the cause of the decrease in certain consitutive enzyme activities (4.E.3). The distribution Of the phosphofructokinase activity in hepatocytes also differed markedly from that in whole liver. The lower remaining enzyme activity in the hepatocytes was found to remain in the superna- tant when the homogenate was centrifuged at 10,000 x g for 5 min. This is in marked contrast to results with whole liver, where most of the homogenate phosphofructokinase activity sediments at low (1000 x g) centrifugal fields (Chapter 3). We observed previously that the particulate phosphofructokinase from whole liver became graduallly solubilized when the particulate material was diluted in a large volume of buffer. This suggested that the binding may be in equilibrium and may depend on the phosphofructokinase concentration. The lower phosphofructokinase activity in the isolated hepatocytes may have caused the enzyme to become solubilized, and may be the reason we observed the enzyme in the soluble fraction. 3. Effect of Glycolytic and Gluconeogenic Incubation Conditions on Phosphofructokinase Activity in Pig Hepatocytes. We were next intersted in determining whether the level and subcel- lular distribution of the pig hepatocyte phosphofructokinase might be affected when the cells are incubated under conditions which promote either glycolysis or gluconeogenesis. In order to observe maximum effects, we used relatively high concentrations of substrates together with hormones in the incubation buffers. Incubation experiments and enzyme assay procedures were performed as described in Materials and 159 Methods (4.0.3). Enzymes were assayed under saturating conditions, to give total activity. The results of a typical incubation experiment are shown in Table 7. When cells were incubated in the presence of 40 mM glucose and 175 AM insulin, the phosphofructokinase activity increased by about 23% above the control (incubated in Krebs-Henseleit buffer with no addi- tions). In the presence of 5 mm glucose, the activity still increased, but only by about 15%. In the presence of 20 mM lactate and 100 AM glucagon, phosphofructokinase activity predictably decreased, by 15%. However, the lactate dehydrogenase activities were found to be unaf- fected by the glucose or lactate treatments. The total phosphofructo- kinase activities in these experiments was still only about 10% of that in whole liver homogenates, and as we found in the experiments describ— ed in the previous section, the activities were found in the soluble fraction. Since the phosphofructokinase activites were assayed at saturating conditions, the variations in the activity must have been due to either changes in the Vmax or in the concentration of active enzyme. 160 Table 7. Effect of Glycolytic and Gluconeogenic Incubation Conditions on Total Phosphofructokinase Activity in Pig Hepatocytesa. TOTAL PHOSPHOFRUCTOKINASE ADDITIONS T0 BUFFER ACTI ITY (mUNITS/IO CELLS) CONTROL (N0 ADDITIONS) 3.98 5 mM GLUCOSE 4.58 40 mM GLUCOSE + 4.88 175 uM_INSULIN 20 mm LACTATE + 3.39 100 uM_GLUCAGON aHepatocytes were isolated and incubation experiments were carried out as described in Materials and Methods. The incubation time was 20 min. For additional explanation, see text (4.0.3). E. DISCUSSION 1. Possible Reasons for Differences in Phosphofructokinase Solubility in Hepatocytes versus Whole Liver. The major differences in our results with isolated pig hepatocytes and those with whole liver (described in Chpater 3) were that the phosphofructokinase activity in hepatocye homogenates was 5 to 15 times lower than in whole liver homogenates, and was soluble, rather than sedimenting at low (100 x g) centrifugal fields. (Obviously there is a large decrease in phosphofructokinase activity during the hepatocyte isolation procedure.) In previous studies with whole liver, the membrane-bound phospho- fructokinase appeared to be in equilibrium with the soluble enzyme, and was gradually solubilized when the membrane fraction was washed repeat- edly or diluted in large volumes of buffer. Therefore it seems probable that the solubility of the hepatocyte phosphofructokinase was due to dissociation of the enzyme from the membrane, due to the lower phoshofructokinase concentration. Our original goal in these studies was to investigate the possible physiological significance of the reversible membrane binding phenome- non of phosphofructokinase. However, the results of the membrane bind- ing study in hepatocytes are inconclusive, since most of the phospho- fructokinase appeared to be lost during the hepatocyte isolation proce— dure. Therefore these results do not exclude the possibility that the 161 162 reversible membrane binding phenomenon of pig liver phosphofructokinase might ocur i_ vivo. Previous studies (Chapter 3) suggested that phosphofructokinase might bind to plasma membranes, and previous workers (102) have suggested that perfusion of liver with crude collagenase might alter the hepatocyte plasma membranes. However, the changes (if any) would probably not affect the inner membrane surface, where we would expect phosphofructokinase would most likely bind. One possible area for further investigation could invovle determin- ing conditions which could minimize the decrease in phosphofructokinase which occurs during the isolation of pig hepatocytes. Several possi- bilities will be discussed in a subsequent section. 2. Contrast of Results of Incubation Experiments with Results of other Workers. The total phosphofructokinase in the incubation experiments (Table 7 and Results) varied in a predictable manner. Under conditions promoting glycolysis (high glucose + insulin) the hepatocyte phospho- fructokinase activity increased, while under conditions promoting gluconeogenesis (high lactate + glucagon), the activity decreased. These results seem somewhat similar to those of Pilkis gt_al. (98). Van Schaftingen and Hers (114), Castano gt_al. (115), and Kagimoto and Uyeda (116). In their incubation studies (with rat hepatocytes), they found the phosphofructokinase activity regulated by changes in the concentration of fructose-2,6-P2 (97), a potent activator of the enzyme. Fructose-2,6-P2 lowers the Km of phosphofructokinase for fructose-6~P (97), and therefore in order to see an activating effect, 163 the enzyme must be assayed at sub-saturating conditions. However in our studies, phosphofructokinase was assayed at saturating conditons, where effects of furctose-2,6-P2 would not be observed. Therefore, the changes in phosphofructokinase activity which we observed were most likely due to changes in the concentration of active enzyme, or in the V max Binkley and Richardson (117) observed that in a certain strain of mice, fasting for 24 hr resutls in a marked decrease in total liver phosphofructokinase activity. In mice receiving insulin injections, the total phosphofructokinase activity increased. When the liver extracts from the fasted mice were treated with glutathione or other thiols, the phosphofructokinase activity was restored to normal. Thus it was necessary to omit thiols from the homogenization and assay buf- fers in order to see the effect. In our studies, we included thiols (DTT and mercaptoethanol) in the buffers, so the effects we observed may be different. However it is also possible that we did not allow enough time for activation to occur. Generally, our observations seem similar to those of Binkley and Richardson (117) in that the phospho- fructokinase activity appears to be regulated by changes in the concen- tration of active enzyme, rather than in the affinity for fructose-6-P. 3. Possible Areas of Future Research on Pig Hepatocyte Phosphofructokinase. a. Possibilities for Improving Isolation Procedure for Pig Hepatocytes. Using recently developed techniques for isolating hepatocytes from large animals (102, 103, 109, 111) we were able to obtain reasonable 164 amounts of pig hepatocytes (about 108 cells from 40 g of liver). The viability of the preparations were sometimes high (up to 80%, by trypan blue exclusion); however, occasionally preparations had to be abandoned due to low viability. One area for future research could be to determine how to reproducibly obtain highly viable preparations of pig hepatocytes. Low hepatocyte viability probably results from cell membrane damage during the mechanical dispersement step after the collagenase perfusion. Adult pig liver appears to contain a high proportion of collagen, and therefore it is especially important to maximize the collagenase activity and perfusion efficiency, in order to obtain lots of dispersed hepatocytes. In an earlier review on rat hepatocyte isolation, Seglen (108, 131) reported several thorough quantitative studies on variables affecting the dispersal of rat liver into intact hepatocytes. Several of the Optimizing conditions described by Seglen for isolating rat hepatocytes might also improve the pig hepatocyte isolation procedure. In addition to giving a review on the deveIOpment of hepatocyte isolation (108), the paper also gives a detailed description of the perfusion apparatus and isolation procedure. Some key points about the procedure which would probably be important in isolating pig hepatocytes are as follows: (1) Collagenase requires Ca2+ ions for activity, and the Optimum concentration was found to be 5 mM. Mg2+ should not be included in the collagenase buffer since it competes with Ca2+ and is inhibitory. 165 (2) The optimum pH for collagenase is about 7.5. Because the liver produces acid during perfusion, the choice of an efficient buffer is important. Seglen found that a bicarbonate/002 buffer system sup- plemented with HEPES (108) gave the best bufering capacity. (3) The optimum collagenase concentration was found to be 0.5 mg/ml. (Hyaluronidase was ineffective, and actually was slightly inhibitory.) Seglen used about 5 ml of the collagenase buffer per gram of rat liver. This ratio should probably be increased by at least 2-fold for pig liver. (4) Perfusion was most efficient using a high flow rate (40 to 60 ml/min). However the perfusion pressure should not distend the tissue, since this might damage cell membranes. Perfusion efficiency was also improved by pre-centrifuging and filtering the collagenase solution to remove undissolved particles, and by including a bubble-trap and filter in the perfusion system. b. Possibilities for Preventing Decrease in Phosphofructokinase Activity in Pig Hepatocytes during Isolation Procedure. We found that a large decrease in pig liver phosphofructokinase activity appears to occur during the hepatocyte isolation procedure. For future investigations of the phosphofructokinase membrane binding phenomenon, (which appears to be concentration-dependent), it would be useful to be able to obtain freshly isolated pig hepatocyte prepara- tions with a higher level of phosphofructokinase activity, nearer to the levels found in whole liver. Previous workers have also observed decreases in other constitutive enzyme activities in isolated hepatocytes. In some cases the decreases 166 appeared to result from leakage of the enzymes through the cell membranes (102, 108). However the protein leakage was probably due to some of the cells having suffered plasma membrane damage from the isolation procedure. Cells with intact plasma membranes, (judged by trypan blue exclusion), would not be expected to leak proteins. Isolated hepatocytes have been found to be freely penmeable to amino acids (108, 118), and show evidence of protein degradation when incubated in substrate-free media (108, 113). The decrease in phosphofructokinase activity we observed may have also been due to protein degradation. Another possibility is that the phosphofructo- kinase may have gone to an inactive disulfide form (117). However this does not seem as likely since our hepatocyte extracts contained dithiothreitol, which probably would have been able to re-activate the enzyme (117). Several investigations have been reported where hepatocyte enzyme activities and metabolic processes have been preserved by including metabolites and hormones in the perfusion and incubation media (108). (For example: hepatocyte glucokinase activity can be preserved by insulin (125); gluconeogenesis from lactate is markedly stimulated by lysine (126), and to a lesser extent by other amino acids (108), NH4+, and K+ ions (127, 132); gluconeogenesis from histidine can only occur in the presence of added methionine (128). In order to preserve the pig hepatocyte phosphofructokinase activity, we would predict that adding glucose (or fructose) to the perfusion buffer would have the biggest effect. Previous workers (123, 129, 130) found that relatively high (50 mM) concentrations of glucose were required in order to promote glycogen synthesis in isolated rat 167 hepatocytes. Glucose concentrations of 20 mM were ineffective. Fructose, however, was effective at promoting glycogen synthesis at lower concentrations (20 mM) (129). These results suggest that high concentrations of glucose (or fructose) would be required in order to preserve hepatocyte glycolytic enzyme activities. This is probably more important for pig liver, which must be perfused for a longer period of time with the collagenase buffer compared to rat liver. Alternatively, supplementing the perfusion media with amino acids might prevent loss of constitutive enzymes from protein degradation. It might also be possible to use a complete cell culture media for the perfusion; however care must be taken to omit compounds that could inhibit collagenase activity, such as Mg2+ and sulfhydryl com- pounds. Perhaps adding vitamins, hormones, and various salts might also help preserve constitutive enzyme activities in pig hepatocytes. CHAPTER 5 STUDIES ON RAT AND RABBIT LIVER PHOSPHOFRUCTOKINASE 168 A. ABSTRACT The subcellular distribution of rat and rabbit liver phosphofructo- kinases was studied, in order to determine whether rat or rabbit hepa- tocytes would be useful for studying the phosphofructokinase membrane- binding interaction. The results of this characterization, and some additional properties of rabbit liver phosphofructokinase, are presented. In homogenates of fresh rabbit liver, in 0.25 M_sucrose, nearly all Of the phosphofructokinase activity sediments at 120,000 x g, with the microsomal fraction. In rat liver however, the phosphofructokinase remains in the soluble fraction. Sucrose gradient velocity sedimenta- tion and flotation studies suggested that rabbit liver phosphofructo- kinase is not actually bound to microsomal membranes. The enzyme also is not bound to particulate glycogen. Rabbit liver phosphofructokinase is therefore probably a large, highly polymerized protein. The enzyme does not dissociate when incubated in the presence Of substrates, products, or related metabolites. The protein-protein interaction is also stable in the presenace of 1% Triton X-100 and 0.2 M_KCl. We conclude that the rat and rabbit liver phosphofructokinases do not have the same membrane-binding property as the pig liver enzyme. 169 B. INTRODUCTION Our initial goal in this study was to investigate the subcellular distribution of phosphofructokinase in rat and rabbit liver, and eventually study the phosphofructokinase-membrane interaction in hepa- tocytes from these animals, since hepatocytes can be isolated more easily and with greater viability fran the rat and rabbit than fran the pig. These studies were done before those described in Chapter 4, and it was found that rat and rabbit liver phosphofructokinases do not have the membrane-binding properties that we observed in pig liver. The studies on pig hepatocytes in Chapter 4 were done subsequently. Initial differential centrifugation studies showed that rabbit liver phosphofructokinase sediments with the microsomal fraction. Further studies were done to determine whether the particulate enzyme was actually bound to microsomal membranes, whether it could be a large, highly polymerized protein, or whether it was associted with particulate glycogen. Other studies were done to determine the effects of substrates, products, and several other treatments on the particu- late rabbit liver phosphofructokinase. 170 0. MATERIALS AND METHODS Most of the experimental procedures were the same as those described in Chapter 3. Additional procedures will be described in detail. 1. Tissues and Reagents. All chemicals were reagent grade. Livers were obtained from rats or while New Zealand rabbits. Animals were starved 12 hr before sacrifice. Rabbits were anesthetized by injection of pentobarbital through the ear vein. Livers were placed on ice after removal fran the animal. The gall blader was removed carefully in order to avoid contamination from bile. 2. Preparation of Liver Homogenates and Differential Centrifugation. Livers were washed and homogenized as described previously (3.0.3). The procedure was started within 10 minutes after removing the liver from the animal. The homogenization buffer was also the same as de- scribed previously (50 mM imidazole°HCl, pH 7.4, 0.25 M_sucrose, 5 mM dithiothreitol). Four ml buffer per gran of tissue was used for homo- genization. The standard buffer, used in all procedures after homoge- nization, contained the same components as the homogenization buffer, plus 3 mM M9012, and for most experiments with rabbit liver phospho- fructokinase, ATP was also included, at a concentration of 5 mM. 171 172 Differential centrifugation was done as described in Chapter 3 (3.0.4) and the legend for Figure 14. 3. Separation of Microsome and Glycogen Fractions. The 120,000 x g pellet fron the differential centrifugation proce- dure consisted of two layers; a hard clear pellet at the bottom of the tube (identified as glycogen by previous workers (80) ), and a reddish layer of material on top, containing the microsomes. The reddish pel- let was removed with a few ml of buffer (without ATP), and the clear glycogen pellets were resuspended in buffer using a Dounce homogenizer. The microsomal fractions were combined and resuspended in a Teflon- glass homogenizer, with additional buffer, and centrifuged again at 120,000 x g as before. The washed microsomal pellet was then resus- pended in the standard buffer (with our without ATP, depending on experiment), using 0.25 ml buffer per initial gram of liver. 4. Sucrose Step-Gradient Flotation Procedure. Aliquots of the washed microsome fraction were mixed with concen- trated (2.4 M) sucrose (in standard buffer) to give a final sucrose concentration of 1.6 or 1.9 M. Aliquots (2 to 4 ml) of this suspension were then transferred to 1/2 x 2" cellulose nitrate tubes (Beckman), overlayed with buffer, and centrifuged in the SW 50.1 rotor at 45,000 RPM for l to 2 hr at 4°. After centrifugation, the upper liquid layers were removed with a Pasteur pipet; floating material at the heavy sucrose interface was collected with a wide-bore pipet and dispersed in 2 to 3 volumes Of buffer. The clear solution below the floating 173 material was collected separately, and the pellet was resuspended in 1 ml of buffer, in a Teflon-glass homogenizer. 5. Determination of Particulate and Soluble Rabbit Liver PFK Activities in Incubation Mixtures. One ml aliquots of the washed microsome suspension (without ATP) were placed in 1/2 x 2" cellulose nitrate tubes. Tubes were then incubated for various amounts of time in the presence of various compounds (metabolites, salt, detergent). Tubes were then centrifuged in the SW 50.1 rotor at 45,000 RPM for 20 min, at 4 or 20°. Supernatants were then collected with Pasteur pipets and pellets were resuspended in 1 ml of buffer. Supernatant and pellet fractions were then assayed for soluble and particulate phosphofructokinase activity, respectively, as described previously (3.0.7). 6. Sucrose Density Gradient Centrifugation. Linear 15 to 40% (w/v) sucrose gradients (in the standard buffer) were formed in 1/2 x 2" cellulose nitrate tubes. Aliquots of washed microsome suspension were diluted with an equal volume of standard buffer with sucrose omitted. (Dilution of the sucrose concentration was necessary to prevent the microsomes from sedimenting while standing at l x g.) Aliquots (100, 200, or 300 pl) of the diluted microsome suspension were layered on top of the sucrose gradients and centrifuged in the SW 50.1 rotor at 47,000 RPM (200,000 x g) for 75 minutes, at 4°. About 35 fractions (5 drops each) were then collected by puncturing the tube. Phosphofructokinase activity in the fractions was assayed using the standard procedure, and active fractions were combined. The 174 presence of microsomes was monitored by measuring the ultraviolet absorbance at 295 nm. Although microsomes absorb more strongly at lower UV wavelengths, the ATP in the buffer also absorbs in the UV. However ATP does not absorb significantly at 295 nm or higher. D. RESULTS 1. Relative Distribution of Rat and Rabbit Liver PFK in Fractions Obtained by Differential Centrifugation. Liver from white New Zealand Rabbits was fractionated by differen- tial centrifugation as described in Materials and Methods. The total phosphofructokinase activity in each fraction was determined by multi- plying the specific activity by the volume of the fraction. The results of a typical fractionation, shown in Table 8, indicate almost all of the phosphofructokinase activity is in the 120,000 x g (microso- mal) pellet. The total recovery of activity was greater than that in the starting homogenate, giving a percent recovery greater than 100%. The 5% activity in the 10,000 x g pellet was eluted with one buffer wash, showing this activity was probably due to some included superna- tant. Most of the 23% activity in the 120,000 x g (post-microsomal) supernatant was found to sediment when this fraction was centrifuged again at 120,000 x g for an additional 0.5 to 1 hr. In 4 different experiments with whole rabbit livers (average weight = 74 g), the homogenates contained about 54 units of phosphofructokinase. All or most of this activity sedimented at 120,000 x g. The specific activity of the microsomal particulate enzyme was about 70 mU per mg protein. (Protein was measured according to the Lowry procedure (29), using bovine serum albumin as a standard.) 175 176 In a differential centrifugation experiment with fresh rat liver, virtually all of the phosphofructokinase activity was found in the soluble fraction (120,000 x g supernatant). In another experiment, the differential centrifugation procedure was done with frozen rabbit liver. Although some of the phosphofructo- kinase activity (22%) was found in the microsomal fraction, most of the TABLE 8. RELATAIVE DISTRIBUTION OF RABBIT LIVER PHOSPHOFRUCTOKINASE IN FRACTIONS OBTAINED BY DIFFERENTIAL CENTRIFUGATION. PERCENT OF INITIAL HOMOGENATE PFK ACTIVITYa FRACTION FRESH LIVER FROZEN 10,000 x g PELLET 5b 0 120,000 x g PELLET 102c 22 120,000 x g SUPERNATANT 23d 62 aThe initial homogenate activity was normalized to 100%. bMost of this activity was removed in the first wash. CThis fraction apparently contained more activity than the starting homogenate, hence the recovery is greater than 100%. dAt least half of this activity was found to sediment if the fraction was re-centrifuged at 120,000 x g for an additional 0.5 to 1 hr. activity was in the post-microsomal supernatant. Also the percent recovery in frozen tissue was lower (84%) suggesting perhaps some inactivation had occured. These results Show the importance of using fresh tissue when studying the subcelluar distribution of the enzyme. 177 2. Location of Rabbit Liver PFK Activity in Subfractions of the 120,000 x g,Pellet. The 120,000 x g pellet obtained from differential centrifugation contained two separate layers, a reddish material on top, and a hard transparent pellet underneath it. Previous workers have found that the transparent material is made up of glycogen and the reddish material contains the microsomes (80). When the two layers were collected separately and assayed, as described in Materials and Methods (5.0.3), all of the phosphofructokinase activity was found in the top layer, while the clear pellet contained no significant activity. This shows that rabbit liver phosphofructokinase is associated with the microsomal fraction, and is not bound to glycogen. 3. Sucrose Step-Gradient Centrifugation of Rabbit Liver Particulate Phosphofructokinase. Aliquots of washed rabbit liver microsome fraction were made 1.6 or 1.9 M_in sucrose, overlayed with buffer, and centrifuged, as described in Materials and Methods (5.0.4). In these experiments, most of the microsomal material floated, and formed a reddish turbid band at the interface between the buffer and the heavy sucrose. The liquid below the turbid band still had a slight turbidity, which remained in solu- tion even when centrifuged for 2 hr. The turbid band and the liquid below it were collected separately and assayed as described previously. All of the phosphofructokinase activity was found in the clarified liquid, while the turbid band contained only a trace of activity, prob- ably carried over from the liquid phase below it. This same result was obtained with both 1.6 and 1.9 M_sucrose in the initial suspension, and 178 in the presence and absence of 5 mM ATP. This shows that rabbit liver particulate PFK must have a density equal or greater than that of 1.9 M sucrose (density = 1.246 g/cm3). These results differ from those with the pig liver particulate PFK, which floated in a sucrose concentration of 1.46 M_(density = 1.190 g/cm3). 4. Stabilization of Rabbit Liver Particulate PFK Activity by ATP. We observed that the phosphofructokinase activity in the rabbit liver microsomal fraction gradually decreases during storage at 0 to 4° (Figure 27). In a preliminary experiment we found that adding ATP to the buffer prevented inactivation of the enzyme. Figure 27 shows the effect of storing the enzyme in the presence and absence of 5 mM ATP (at O-4°) on the stability of the microsomal phosphofructokinase. In the absence of ATP, the phosphofructokinase gradually decreased over several days, and little or no activity remained after 8 days. In the presence of 5 mM ATP, the activity still decreased, but at a much slower rate; after 8 days the activity stabilized at 42% of the initial activity. This 42% activity remained stable for 4 days, and then gradually decreased. In another experiment (Figure 27), an aliquot of washed microsomes was incubated without ATP for 4 days, at which time the activity had decreased to 22%. ATP was then added, to a concentration of 5 mM. As shown in Figure 27, some reactivation had occured, bringing the activ- ity to 34%. This activity was stable for several days. Previous workers (119) have found that the purified rabbit liver phosphofructokinase undergoes a transition from an active to a less active form when incubated for 2 min (at 25°) in the absence of ATP. 179 .25 m 00 :000000c00:00 0 00 .00>00: 0:0 00 0:0 00 00000 003 00< 0:0 .000: :0 0000>00 003 A00< 0:0:003V 0000000 00:0:00 0:0 .60 00 :0000000:0 0000 0 L000< .A0.o.mv 000>L00:0 0500 0000000:0 0:0 00 00:05:0000 003 x00>0000 000:0:0000000:000:0 .23 m 00 :000000:00:00 00:00 0 0>0m 00 00000000 0:0 00 0:0 00 00000 :0:0 003 A:000:_00 x0000.z _.ov 00< .00000000 N 00:0 0000>00 003 .00< 0: :003 0000:: 0:00:000 0:0 :0 .:000:00000 050000005 0:0 .60 00 0:00 0:03 0:0000000:H .Am .N.o.mv 000:005 0:0 0000:0002 :0 0000:0000 00 0000000: 0:03 005000000: L0>00 000:0: .0H< >m >HH>HFQ< mmH: HHmm<0 20 no I ............. 00:2 :2 as m T ................ .Juuax / 000000000000 4 0000000000 . . 1101110 , /0 , a: 02: v OIJOKPZOQ :2 2.: 0\« x .. 0 0/ I , cm 80 All/111317 HSVNINOLOTIEHOHdSOHd BAIIV'IEIEI 181 Adding ATP to the incubation mixture resulted in reactivation of the enzyme, showing the process is reversible. The degree of association of both forms of the enzyme appeared to be unchanged, and thus the effect was attributed to a reversible conformational change of the enzyme. Our results show that ATP also prevents a complete inactivation of the enzyme during long term storage. The slight activation we observed when 5 mM ATP was added after a 4 day incubation (at 0-4°) may have been due to conversion of the remaining less active fonn of the enzyme to the active fonn. 5. Effect of Metabolites, Detergent, and KCl on Particulate Rabbit Liver Phosphofructokinase. The next question we asked was whether the particulate rabbit liver phosphofructokinase could be solubilized by treatments which solubilize the particulate pig liver enzyme. Solubilization experiments were done with 1 ml aliquots of the washed microsome fraction, as described in Materials and Methods (5.0.5). ATP was not included in the standard buffer. The treatments tested for solubilization activity were: 6 mM concentrations of F-6-P, F-1,6-P2, G-6-P, AMP, ADP, and ITP, 9 mM ATP, 1% Triton X-100, and 0.2 M_KCl. The control contained no added compounds. After incubating at O or 23° for time periods of 5, 30, or 60 minutes, the suspensions were centrifuged at 190,000 x g for 15 min. In all tests, including the control, all of the phosphofructokinase activity was found in the pellet fractions, and therefore no solubiliz- ation had occured from any of the treatments. Metabolite determina- tions (3.0.8) on the supernatant fractions showed that the added 182 metabolites were not depleted during the incubations. In the suspen- sion treated with Triton X-100, some solubilization of microsomal mate- rial had occured, shown by an increase in turbidity of the supernatant and a decrease in pellet size. However all of the phosphofructokinase activity was still found in the pellet. These results differ markedly fron the results with the pig liver phosphofructokianse (3.0.1, 2, 3), where we found the enzyme was rapidly solubilized by many of these treatments (Table 4.) 6. Sucrose Gradient Sedimentation Velocity of Rabbit Liver particulate Phosphofructokinase. A major question we wanted to answer was whether the phosphofructo- kinase in the microsomal fraction was bound to the microsomal mem- branes, or whether the enzyme was polymerized to form a high molecular weight particle. The sucrose step-gradient flotation experiments show- ed that phosphofructokinase did not float with most Of the microsomal fraction in heavy suscrose, suggesting that the enzyme was not bound to microsomes. Another approach we next used was sucrose density gradient sedimen- tation velocity. Aliquots of the washed microsome fraction were applied on 10 to 40% sucrose gradients and centrifuged as described in Materials and Methods (5.0.3). The results of a typical experiment are shown in Figure 28; after centrifugation a single diffusely turbid band was Observed, about 3/4 of the way into the gradient. This band cor- responded to the 295 nm absorbance peak (Figure 28), and was attributed to the major microsomal fraction, since it was the only turbid band observed. However the phosphofructokinase activity was found in a peak 183 .005000:0we 0000000 0:0 00 0000000 F: 000 0 :003 .omm 00 :0: 003 0:050:00x0 0:0 .Ao.u.mv 000:00: 0:0 0000:0002 :0 0000:0000 00 00:00:0000 0:03 >00>0000 000:0:000::00:000:0 0:0 00:00:0000 >2 0:0 .0000:00000:0 .::: 0:03 00:0000:m 000:0:m .Am.u.mv 000000>0:0 0000:0000 00 00:000:0 0:03 005000:00E :0>0_ 000:0: .onhHHoo:m> HzmHom .mmzomomqu 20mm mmH4 HHmm000< :0 000: mmmxz mAmHu 4~p<4mm mz>sz omHmHmaa 6.0N:-:._-60606::0 :0 3000000000 :0: :.:.60606::: .:0< 666:0:66: 6:6: 606:0:66: 6:6: 00:0:00: 000000:<00: .. .. >0:>:00< 6.0:: 0.0 0.6:: 0: - m 606:_::6: 6:6: zH<0ZHHmmaom0 .wHa oz< .HHmmH: no mmHHmmaomm zH mmozmmmmmHa mod