1984a, 1984b; Kayne and Price, 1972,1973; Kwan and Davis,. 1980, 1981). ..... form of an increase in absorbance at 300 nm at low denaturant concentrations ...
Vol. 263, No. 6, Issue of February 25, pp. 2787-2793, 1988 Printed in U.S.A
THEJOURNALOF BIOLOGICAL CHEMISTRY 0 1988 hy The American Society for Biochemistry and Molecular Biology, Inc.
Domain Interactionin Rabbit Muscle Pyruvate Kinase I. EFFECTSOFLIGANDSONPROTEINDENATURATIONINDUCED
BY GUANIDINEHYDROCHLORIDE* (Received for publication, July 2, 1987)
Thomas G . Consler and James C. Lee From the E. A. Doisy Department of Biochemistry, St. Louis University School of Medicine, St. Louis, Missouri 63104
change on pyruvate kinasehave been documented by kinetic, equilibrium binding, and structural data (Oberfelder et al., 1984a, 1984b; Kayne and Price, 1972,1973; Kwan and Davis, 1980, 1981). The simplest model which can consistently incorporate all the experimental data reported by Oberfelder et al. (1984a, 198413) is a concerted allosteric model (Monod et al., 1965). One of the major observations of this study is that, upon bindingof Phe to oneof the four identical subunits, the effector causes all subunits to change conformation, as monitored by difference sedimentation velocity and equilibrium binding studies (Oberfelder et al., 1984a). At present, the specific features of conformational change are not known, although they may involve the three domains revealed by x-ray crystallographic data (Stammers and Muirhead, 1975, 1977; Stuart et al., 1979). Intersubunit contact occurs between domains A and C of adjacent subunits and between adjacent A domains as well as adjacent C domains. The active sitelies in thecleft between the A and B domains. There are no functions ascribed to domainsB and C, although a potential secondarynucleotide-binding site between domains A and C has been proposed.Another interesting feature is that thereis more flexibility in domain B than in the other domains. Since the binding of Phe perturbs the kinetic behavior of the enzyme, it must perturb the active site; and since Phe is not a strict competitive inhibitor with respect to the substrate, Phe must bind at a site other than the active site.Thus,binding of Phemust trigger some structural changes that are propagated throughthe molecule. These Being an important regulatory glycolytic enzyme, skeletal structural changes may represent local changes surrounding muscle pyruvate kinase has been studied extensively. Much the ligand-binding sites or long-range perturbations propainformation is known about the kinetic mechanism (Mildvan gated through changesin domain interactions which may lead andCohn, 1965, 1966; AinsworthandMacFarlane, 1973; to changes in the stability of the protein. In order to gain more insight into the solution behavior of pyruvate kinase DannandBritton, 1978; Hassett et al., 1982), andsome aspects of its regulatory behavior have been examined. The and tocorrelate this behavior with the crystal structure, this enzyme catalyzesthe reaction: P-enolpyruvate + ADP + study was undertaken. The major objective of these experiidentified by xpyruvate ATP, and requires a divalent cation, preferably ments was to determine whether the domains M$+, and a monovalent cation,K’, for activity (Boyer, 1962). raycrystallography would behave as independent domains I n uiuo, the production of ATP plays a major role in the that could be detected insolution. energy fluxof the cell. Thus, it is natural that pyruvate kinase EXPERIMENTAL PROCEDURES holds a key position in theglycolytic pathway, and its activity can be expected to be subjected to a pattern of regulation. LMaterials Phenylalanine (Phe)’ has beenshown by steady-state kinetic Pyruvate kinase from rabbit muscle in 3.2 M (NH4),S04suspension, datatoexertanallostericinhibitiononpyruvatekinase trypsin, and tricyclohexylammonium salt of P-enolpyruvate were (Carminatti et al., 1971). Although the in uiuo significance of purchased from Boehringer Mannheim. Pyruvate kinase was greater Phe as an allosteric effector is unclear, itseffect on the kinetic than 98% pure as monitored by SDS-polyacrylamide gel electrophobehavior and concomitant induction of conformational resis and had a specific activity >250. Tris base, Tris-HC1, disodium
The structural stability of rabbit muscle pyruvate kinase was examined. The unfolding of pyruvate kinase was induced by guanidine hydrochloride, and the process was monitored by spectroscopic techniques (fluorescence and UV absorption) and hydrodynamic measurements (sedimentation velocity, sedimentation equilibrium, densimetry, and viscometry). The spectroscopic techniques revealed that the unfolding of pyruvate kinase induced by guanidine hydrochloride is not a simple cooperative process. This suggests that different regions of pyruvate kinase are unfolding with different efficiencies in response to the denaturant. These regions are most likely related to the domain structures observed by x-ray crystallography. In the presence of L-phenylalanine, the allosteric inhibitor, the denaturation process becamemore cooperative, and the enzyme dissociated and unfolded at a higher denaturant concentration. The binding of phenylalanine also induced a structural change in the enzyme, rendering it more susceptible to tryptic digestion. One of the peptides, the production rate of which was increased, was isolated and sequenced. Its N terminus is located at the interface between two domains, one of which contains the active site.This evidence indicates structural changes, probably involving domain-domain interaction, for pyruvate kinase in response to phenylalanine binding.
+
* This work was supported in part by National Institutes of Health Grants NS-14269 and DK-21489. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The abbreviations used are: Phe, L-phenylalanine; SDS, sodium dodecyl sulfate; GdnHC1, guanidine hydrochloride.
salt of ADP, NADH, lactate dehydrogenase type 11, and Sephadex G25 were all obtained from Sigma. Phe and ultrapure GdnHCl were purchased from Schwarz/Mann and Heico, respectively. Acrylamide was obtained from Bethesda Research Laboratories.
Methods Rabbit muscle pyruvate kinase was desalted using a two-step process. The ammonium sulfate precipitate was concentrated by
2787
2788
Structure Solution
centrifugation at 10,000 rpm in a Sorvall SS 34 rotor for 15 min, and the pellet was resuspended in a minimal volume of TKM buffer (50 mM Tris, 72 mM KCI, and 7.2 mM MgSO, (pH 7.5)). The desalting process involved passage through a "dry" Sephadex G-25 column to minimize dilution, followed by filtration in a standard Sephadex G25 column to remove residual salt. It was found that both steps were necessary in order to desalt completely the protein preparation. The protein concentration was determined by measuring the absorbance at 280 nm, using a molar absorptivity of 0.54 ml/(mg.cm) (Boyer, 1962). GdnHCl was prepared as a 6.0 M stock solution in TKM buffer, and all experimental solutions were prepared from this concentration of denaturant by dilution with TKM buffer. GdnHCl solutions were prepared at aconcentration two times that desired for the final experimental condition so that, upon mixing with an equal volume of protein solution a t 1.0 mg/ml, a final solution wouldbe obtained containing 0.5 mg/ml pyruvate kinase and the predetermined final concentration of GdnHCl. For experiments otherthan the difference ultraviolet absorbance measurements, the pyruvate kinase and GdnHCl solutions were mixed and allowed to equilibrate for a t least 1 h before measurement. The samples for UV spectroscopy had to be mixed within the cuvette so that base-line spectra could be obtained. Experimental samples that involved the allosteric ligand, Phe, were prepared by adding this ligand to thestock protein solution in such a manner that the 1.0 mg/ml pyruvate kinase solution was made 30 mM in Phe. This ensured that the final concentration of Phe would be 15 mM and that the protein would be pre-equilibrated with the ligand before measuring the denaturation. Denaturation of Pyruvate Kinase Spectroscopic Methods-The denaturation of pyruvate kinase was monitored by the change in its ultraviolet spectrum using a Cary model 118 spectrophotometer. Samples were placed into matched quartz tandem cuvettes. Base lines were generated for each sample before mixing the sample cuvette. Samples were allowedto equilibrate before the perturbed spectra were recorded. Equilibration time varied from 15 to 60 min depending on denaturant concentration. Samples were routinely measured also after 24 h, and no differences were observed. The denaturation of pyruvate kinase was also monitored by change in fluorescence intensity using a Perkin-Elmer model 512 doublebeam spectrofluorometer. The excitation and emission wavelengths were295 and 350 nm, respectively. Uncorrected emission spectra were also recorded on selected samples. The sample was excited at either 280 or 295 nm, and the emission wavelength was varied from 300 to 400 nm by scanning this region; the wavelength of the maximal fluorescent emission intensity was determined from these spectra. Sedimentation Velocity-The Beckman-Spinco Model E analytical ultracentrifuge equipped with an ultraviolet scanner was used to measure the sedimentation coefficient of pyruvate kinase as a function of GdnHCl concentration. At low denaturant concentrations ( 4 . 0 M GdnHCl), Kel-F-coated aluminum double-sector centerpieces were used in an An-D rotor and centrifuged a t 60,000 rpm. Scans were taken at 4-min intervals, and the protein boundary was monitored by absorbance at 280 nm. At higher denaturant concentrations, a synthetic boundary cell was used. Solvent was layered onto the protein sample by slowly bringing the rotor up to speed, thus immediately generating a boundary. At intermediate denaturant concentrations (i.e. during the midpoint of the denaturing process), scans were quite complex, often reflecting the existence of multiple species. To analyze correctly these scans, weight average sedimentation coefficients, Sobs, were determined from the centroid. The observed weight average sedimentation coefficients were corrected for solvent conditions using the density, viscosity, and partial specific volumeparameters measured. Sedimentation Equilibrium-Molecular weights were determined by sedimentation equilibrium measurements as a function of GdnHCl concentration using the high-speed procedure of Yphantis (1964). Standard Kel-F-coated double-sector centerpieces were used, and rotor speed was varied from 12,000 to 24,000 rpm depending upon the expected molecular weight of pyruvate kinase under the particular condition. Measurements were made at 0, 0.6, 1.0, 1.25, and 6.0 M GdnHC1. Attainment of equilibrium usually took about 24 h and was assessed by the superimposition of scans taken hours apart. The apparent weight average molecular weights were calculated using the values of density and partial specific volume determined under these experimental conditions.
of Pyruvate Kinase Density Measurements-The apparent partial specific volumes, cp, of pyruvate kinase in TKM buffer in the presence of varying concentrations of GdnHCl were obtained by measuring the densities of protein solutions and those of respective solvents against which they had been dialyzed according to publishedprocedures (Lee et al., 1979). The densities were measured with a precision density meter (Mettler/ Paar DMA-O2D) at 20.00 0.01 'C. The concentration and density data were combined to obtain the apparent values of cpz at each protein concentration, c, according to Equation 1:
*
where p. and p are the densities of solvent and the protein solution, respectively, and c is the protein concentration in grams/milliliter. Viscosity Measurements-The viscosity of solvent and protein solutions was determined using an Ostwald-type viscometer with a flow time of 387 s (at 5.0 "C) or 256 s (at 20.0 "C) for water. The temperature was regulated by a Cannon constant temperature bath. Solvent measurements were used for correctional factors in sedimentation experiments. Protein solution measurements were made in order to demonstrate structural differences in different protein solutions. These measurements were carried out a t 5.0 "C because the protein was found to aggregate and precipitate at higher temperatures in the viscometer. Readings were taken repetitively (5-10 times) and then averaged. Measurements were taken with different protein solutions of varying concentration. Results were expressed as the intrinsic kinematic viscosity, [ q ] , as defined by Equation 2 (Tanford, 1955): [q] =
lim,(t
- tJ/t,c
(2)
where t is the capillary flow time of solution, t, is the flow time of solvent, and c the concentration in grams/milliliter. TIyptic Digestion-Pyruvate kinase was proteolytically digested in the absence or presence of 15 mM Phe, 15 mM alanine, or 7.2 mM MgSO, in Tris buffer containing 72 mM KC1 and 2.0 mM EDTA at pH 7.5 and 37 "C. The ratio of trypsin to pyruvate kinase was 1:200 (w/w). This is roughly a 1:20 molar ratio of trypsin to pyruvate kinase. Aliquots were removed from the reaction mixture at various time points and were mixed with an equal volume of sample buffer (5% SDS, 1% P-mercaptoethanol, 1% dithiothreitol, 50 mM Tris, and 1%bromphenol blue). These samples were boiledimmediately to stop the reaction. Twenty-five-gl aliquots for each time point were loaded per lane in 15% SDS-polyacrylamidegels. Electrophoresis was carried out at 20 mA for about 12 h. Gels were either stainedwith Coomassie Brilliant Blue or alternatively silver-stained by a published procedure (Field et al., 1984)and scanned with a Pharmacia LKBBiotechnology Inc. laser densitometer. Results are expressed as thepercent area of each peak relative to the total of all peaks. Molecularweight standards were included in each gel for determination of the apparent sizes of these peptides. Preparation of Peptides for Sequencing-Fifteen percent acrylamide gels containing trypticdigests were lightly stained with Coomassie Brilliant Blue without trichloroacetic acid fixation. Bands were cut out and equilibrated with 0.05 X TBE buffer (89 mM Tris, 89 mM boric acid, and 2 mM EDTA (pH 8.0)) with 1%SDS overnight. The peptides were electroeluted from these gel slicesusing an ISCO sample concentrating apparatus. Solution containing the electroeluted peptide was then dialyzed overnight against 10 mM phosphate buffer (pH 7.5), and thissample was lyophilized.Remaining SDS andstain were extracted from this lyophilized sample by a solution of acetone/ triethylamine/acetic acid (905:5). One ml of the extraction mixture was added to thelyophilized sample, vortexed, and kept on ice for 10 min. The sample was then centrifuged in an Eppendorf microcentrifuge for 10 min. The supernatant was removed, and the pellet was re-extracted as described. After these extractingprocedures, the pellet was then washed twice with acetone and subsequently resuspended in 10 mM phosphate buffer for sequencing. SequenceAnalysis-The isolated polypeptide was sequenced by automated Edman degradation on an Applied Biosystems 470A protein sequenator. Typically, samples of 0.5-1 nmol were analyzed, and the phenylthiohydantoins were identified by reverse-phase high pressure liquid chromatography analysis on an Altex Ultrasphere OD-SPTH column (Grant et al., 1983). The mobile phase was7.5 mM sodium acetate (pH 5.4), and themodifier was a mixture of methanol/ acetonitrile (173). Thecolumn was run at 34 "C with a constant flow rate of 1.5 ml/min.
Structure Solution
of Pyruvate Kinase
2789
in the spectrum of pyruvate kinasewith increasing concentrations of GdnHCl, as shown in Fig. 2. These changes took the In an effort to elucidate theeffects of ligands on pyruvate form of an increase in absorbance a t 300 nm at low denaturant kinase domains, the stabilityof the enzyme was studied. The concentrations (C0.8 M GdnHC1) and a decrease in absorbdenaturation of pyruvate kinase was followed by change in between 0.5 and 2.0 ance at 288 nm at GdnHCl concentration fluorescentintensitywithexcitationand emission waveM. Since the spectroscopic changes at these two wavelengths lengths a t 295 and 350 nm, respectively. In the absence of were apparently taking place simultaneously, the data were a function of Phe,thechangeinfluorescentintensityas initially analyzed by taking theoverall absorbance changes at GdnHCl concentration can be representedby two phases, as 288 and 300 nm. When the data for pyruvate kinase in the shown in Fig. 1. The initialphase is characterized by an absence of Phe were plotted, the denaturation profile apincrease in therelative fluorescent intensity at GdnHCl con- peared biphasic, as shown in Fig. 3. This again suggested that centrations ranging from0 to 0.8 M. With further increase of the unfolding of pyruvate kinase is not a simple two-state GdnHCl concentration, the relative fluorescent intensity deprocess, and separate domains may be unfolding independcreases and approaches a minimum value at 2.0 M GdnHCl. ently. In the presence of 15 mM Phe, the initial transition These results imply that the denaturationof pyruvate kinase was no longer observed, as shown in Fig. 3. The positive by GdnHCl as monitored by fluorescence does not strictly absorbance at 300 nm was significantly reduced, although the conform to the two-state process. In the presence of 15 mM Phe, the fluorescent data significantly are different, especially I for low GdnHCl concentrations. The initial phase of increasing fluorescent intensity with increasing GdnHCl concentration is still present; but there is only a maximum increase of 28% in signal intensity, as compared with a 72% increase in the absenceof Phe. The wavelength of maximum emission was also examined. 3 In both casesof pyruvate kinase and the pyruvate kinase-Phe complex, the maximum wavelengths were shifted with increasing denaturant concentration. The extent of shift was dependent on the excitationwavelength, 10 nm (from 345 to 355 nm) and 14 nm (from 339 to 353 nm) at the excitation 0.01 wavelengths of 295 and 280 nm, respectively, when the denaturantconcentration was increased from 0.6 t o 3.0 M. Within this concentration range, the shift was monotonic. These shifts to higher wavelengths indicate a change in the local environment of the fluorophores to thatof higher polarity. This observation is expected sincepyruvatekinaseis unfolded by increasing GdnHCl concentration, and fluorothe phores become more exposed to the polar solvent. Another interesting observation is the shift of maximum emission wavelength due to the presence of Phe. In the absence of GdnHCl, the presence of 12 mM Phe. caused a shift of 5 nm regardless of the excitationwavelength. This increase in emisI of Phe to pyruvate 300 sion wavelength indicates that the binding 6(3 40 Wavelength hrn) kinase induces a change in the fluorophore environment to that of higher polarity. FIG. 2. Difference UV spectra induced by GdnHC1. DenatuThe denaturation of pyruvate kinase was also followed by ration was carried out in TKM buffer at 23 "C as a function of difference spectroscopy. Scans of the near-ultraviolet region GdnHCl, and the protein concentration was 0.5 mg/ml. -, 0 M (340 to 240 nm) revealed that there were distinct alterations GdnHC1; - - - - -, 0.7 M GdnHC1; . . . ., 1.0 M GdnHC1; -.-, 1.3 M RESULTS
I
-
1
GdnHC1; -,
3.0 M GdnHCl.
0 8fD
I
OD
'.
I
cmc. of GdnHCl ( M 1 conc of GdnHCI(M)
FIG. 1. Denaturation of pyruvate kinase as monitored b y change in fluorescent intensity. Excitation and emission wavelengths were 295 and 350 nm, respectively. The enzyme concentration was 0.5 mg/ml. Denaturation was carried out in TKM buffer at 23 "C as a function of GdnHCl. 0 and 0, denaturation of pyruvate kinase and enzyme pre-equilibrated with 15 mM Phe, respectively.
FIG. 3. Denaturation of pyruvate kinasebyGdnHCl as measured bydifference UV absorbance. The data are presented as fraction of denatured state, f D , which is monitored by the summation of AA,, and AA2= and has been normalized. 0 and 0, denaturation of pyruvate kinase and enzyme pre-equilibrated with 15 m M Phe in TKM buffer, 23 'C, respectively. The protein concentration was 0.5 mg/ml.
Solution Structure of Pyruvate Kinase
2790
change at 288 nm remained. The denaturationprocess, in the presence of Phe, apparently conforms to a simple two-state unfolding process. Assuming that the extentof pyruvate kinase denaturation is proportional to the change in absorbance at 288 nm, the denaturationcurvesappeartoindicate single transitions, which are affected by the presence and absence of Phe, as shown in Fig. 4. The curve describing the unfolding of the pyruvate kinase-Phe complex is shifted initially to higher denaturant concentration and appears to be more cooperative innature.The curvedescribing theunligatedproteinis skewed, suggesting less cooperativity in unfolding. Since the unfolding process monitored at 288 nm is apparently transparent to local structural changes and probablyreflectsa global change, these denaturation curves were further analyzed in order to describe the thermodynamics of unfolding by assuming that pyruvate kinase undergoes a simple twostate transition. The regions of the denaturation transition that represented the actual change in absorbancewere analyzed as follows:
linear extrapolation procedure of Pace (1975) according to Equation 4: AGD =
+ n (denaturant)
(4)
an extrapolation of the plots of AGO versus GdnHCl concentration to zero denaturant would yield the values of AGoHz0 in aqueous solvent in the absenceof denaturant. The values of AGDHzo are 4 zk 0.3 and 5.2 f 0.3 kcal/mol in the absence and presence of 15 mM Phe, respectively. Hence, there is a difference in the overall transition energy of -1.2 kcal/mol, possibly reflecting the amount of energy by which the tetrameric pyruvate kinase is stabilized as a result of the binding of Phe. In order to characterize the quaternary structure of pyruvate kinase during the denaturation process, sedimentation techniques were employed. For all of the sedimentation data, measured values for the partial specific volume of pyruvate kinase andsolvent density andviscosity were employed inthe calculations. Results from sedimentation velocity studies are shown in Fig. 6. The weight average sedimentation coefficient K = f D / ( l- f D ) = e-"GD:D/RT ( 3 ) is constant from 0 to 0.6 M GdnHC1. Between 0.6 and 1.1M, there is a gradual decrease in the weight average sedimentawhere K is the equilibrium constant, f D isthe degree of tion coefficient, &+, although the presence of 15 mM Phe conversion from the initial to the final state in the two-stateapparently shifts the curve to the right. This shift implies process, AGO is the free energy change inunfolding, R is the that Phe stabilizes the structure of pyruvate kinase, and it gas constant, and T i s temperature. Knowing the equilibrium requires higher concentration of GdnHCl to denature it. At higher denaturant concentration, the value of &,,reaches a constantsat various GdnHClconcentration, AGD canbe minimum of1.9 S, indicating a completedissociation and calculated, and the results are shownin Fig. 5. Using the probably unfoldingof pyruvate kinase. Sedimentation equilibrium was employed to define the molecular weightof pyruvate kinase at different concentrationsof GdnHC1, and the results are shownin Table I. In the absence of Phe, the enzyme remains as a tetramer at 0.6 M GdnHC1; however, at 1.0 M, it is completely dissociated into monomeric form. Yet, even at 1.1M GdnHCl in the presence of 15 mM Phe, pyruvate kinase a tetramer and only maintains its quaternary structure as 0.31 dissociates at 1.25 M as indicated by a weight average molecular weight of 67,000. Hence, sedimentation dataalso revealed that Pheinduces greater stabilityof the enzyme, a conclusion 0 0.5 I.o 1.5 20 3.0 in agreement with the spectroscopic results. The sedimentaconc. of GdnHCl ( M I FIG. 4. Denaturation of pyruvate kinase by GdnHCl as 12 measured by difference U V absorbance at 288 nm. ExperimenI
tal conditions were as described for Fig. 3.
3t 0'
0.5
1.0
1.5
2.0
14.0
conc. of GdnHCl ( M )
FIG. 6. Weight average sedimentation coefficient as a function of GdnHCl concentration. The protein concentration was 0.5 mg/ml, and the sedimentation boundary was monitored by absorption at 280 nm. Symbols are as described for Fig. 3.
TABLE I Apparent weight average molecular weight ( M J of pyruvate kinme I
0
I
04 00 12 conc of GdnHCl (M)
16
FIG. 5. Free energy of denaturation of pyruvate kinase by GdnHC1. AGD was determined as described in the text. Data shown were derived from that represented 230,000 in Fig. 4. A , pyruvate kinase in TKM buffer at 23 "C; B, pyruvate kinase with 15 mM Phe in TKM buffer at 23 "C.
Pyruvate kinase sample GdnHCl conc +Phe
-Phe
M,
M
0.60 1.00 220,000 1.10 1.25
235,000 58,000
67,000
2791
Solution Structure of Pyruvate Kinase tion studies specifically identified that the quaternary structure of pyruvate kinase is stabilized by the presence of Phe. The change in quaternary structure was further probed by determining the intrinsic viscosity of the protein for both of the conditions, i.e. in the presence and absence of Phe at 1.1 M GdnHCl. The results of these experiments are summarized in Table I1 along with data published previously (Doster and Hess, 1981). It is evident that theviscosity differs as afunction of denaturant concentration. The values for [q] are 4.6 and 39.0 ml/g for native and denatured pyruvate kinase, respectively; thus, these values serve to define the two end statesof pyruvate kinase. At 1.1 M GdnHCl, the values of [q] are 36.1 and 18.0 ml/g for the enzymes in the absence and presence of 15 mM Phe, respectively. Hence, these results indicate that the presence of Phe stabilizes pyruvate kinase in a form closer pyruvate to thenative state in1.1M denaturant. Without Phe, kinase appears to be nearly completely unfolded at this concentration of denaturant. All of the physical parameters characterizing pyruvate kinase presented in this report are consistent with the concept that Phe induces structural changes in the enzyme so that the enzyme is stabilized. In an attempt to probe further the solution properties of the enzyme, it was subjected to partial tryptic digestion in thepresence of various ligands. Conditions were adjusted such that theappearance of peptide was linear with time so that ratescould be determined for each particular peptide of interest. There are seven distinctive bands (a-g), one of which is the pyruvate kinase subunit (g), the other six being peptides derived from tryptic digestion of the pyruvate kinase subunit. The two smallest peptides are not separated by the densitometric scan and appear as one slightly skewed band (a), asshown in Fig. 7. The relative intensities of these bands change as a function of time with a progressive accumulation of the bands representing smaller polypeptides. The rates of formation for bands a and b are themost affected by the identity of ligand present. When the tryptic digestion was conducted with pyruvate kinase in the presence of 15 mM Phe or in the absence of M$', the resulting digest rates were fast, as shown in Fig. 7B. On the other hand,when the experiments were conducted with pyruvate kinase in the presence of7.2 mMM$' or both 2 mM P-enolpyruvate and 7.2 mM M2', the digestion rates were at least 4-fold slower, as shown in Fig. 7A. A control experiment was conducted in the presence of 15 mM alanine, which binds to pyruvate kinase competitively with respect to Phe, but does not induce a structural change (Kayne and Price, 1973; Oberfelder, 1982). There is no change in the digestion rates, i.e. they remain slow. The various digestion rates of the enzyme under different experimental conditions are summarized in Table 111. Thus, tryptic digestion seems to be a sensitive monitor of pyruvate kinase conformation. Since the rate of formation of peptide b is significantly increased in the presence of Phe, this peptide was isolated and sequenced. The results are shown in Fig. 8. Comparing this sequence with that for residues 208-222 of both cat and chicken muscle pyruvate kinase (Muirhead et al., 1986; Lonberg and Gilbert, 1983), there was 100% homol-
FIG.7. Densitometric scans of SDS-polyacrylamide gel electrophoresis analyzing tryptic digests of pyruvate kinase. Assay conditions were described under "Experimental Procedures." The protein concentration was 0.5 mg/ml. Time course was 1-6 h. Experiments were conducted in the presence of 7.2 mMMgZ+ ( A ) and 15 mM Phe ( B ) . Both sets of scans were adjusted to the same full scale. TABLE III Summary of tryptic digestion ratesof pyruvate kinase Peptide Condition g
a
b % area/h
15 mM Phe 15 mM Phe + '. 7.2 mM Mg2' 7.2 mM M$' 15 mM Ala + 7.2 mM Mg2+ 15 mM Ala
-10.60 & 1.32 -10.71 -3.05 f 1.320.56 -3.11 0.44 -11.30
& 2.02
3.77 f 0.47 2.90
*
0.32 f 0.02 0.35
0.65 f 0.37 f 0.10
3.90 -C 0.48 3.04
R E S I D E # 206 CAT K-G-V-N-L-P-G-A-A-V-D-L-P-A-V-S-E-K-...
........
2.64 f 0.55 3.94
. .K-G-V-N-L-P-G-A-A-V-D-L-P-A-V-S-E-K-. ... (K-'fV-N-L-P-G-A-A-V-D-L-P-A-V-S-E-v-.
&
1.62
223
CHICKEN.. RABBIT.
I
.. ..
HINGE
I
. I
"
TABLEI1 Intrinsic viscosity of pyruvate kinase at pH7.5 and 5 "C Condition 1.11 mlfg
Native +1.10 M GdnHCl + 15 mM Phe +1.10 M GdnHCl +2.5 M GdnHCl "Data are from Doster and Hess (1981).
4.5" 18.1 36.1 39.0"
_f
FIG.8. Amino acid sequence from peptide b and comparison to other sequences of pyruvate kinase/Ml isozymes. Residue number is labeled above the sequence. Cat and chicken M1 sequences were taken from Muirhead et al. (1986). ogy. Residue 208 is located at the C-terminal part of domain
B, as the polypeptide chain rejoins domain A. Hence, this result implies that at least part of the effect of Phe is to induce structural changes in pyruvate kinase so that some
Solution Structure of Pyruvate Kinase
2792
region(s) of theprotein is more accessible to cleavageby trypsin. At present, there is no direct evidence to indicate that the peptide bond between residues 206 and 207 is the primary tryptic site whose accessibility is influenced by ligands. Trypsin may hydrolyze other peptide bonds first. As a consequence of such hydrolysis, the region containing residues 206 and 207 may becomeexposed, thus leading to an increased rate of release of peptide b. DISCUSSION
One of the issues in allosteric regulation is the mechanism of transmittance of information from one subunit to another in a multisubunit protein. It is even more intriguing in a protein which contains multiple domains, such as muscle pyruvate kinase (Stammers andMuirhead, 1977; Stuart et al., 1979). The functions of these domains in pyruvate kinase need to be defined, and thedynamic interactions among these domains are yet to be explored. In this report, the interaction between domains as reflected by the stability of the enzyme was probed by denaturation in GdnHCl. In addition, the effects of Phe on the interactions between these domains was also studied. The major conclusions derived from this report are that there are regions in pyruvate kinase which exhibit differential stability toward GdnHCl denaturation and that Phe most likely induces changes in domain-domain interaction such that the stability of the enzyme is affected. Let us examine the evidence that support these conclusions. In theabsence of Phe, the denaturation of pyruvate kinase by GdnHCl was monitored by various spectroscopic techniques and at different wavelengths. If the denaturation process were highly cooperative, like most proteins studied, then similar results should be observed regardless of the wavelength or techniques (Ginsburg and Carroll, 1965; Privalov, 1979)employed in monitoring denaturation. However, in pyruvate kinase, different results were obtained depending on the wavelength and spectroscopic technique employed. These observations are consistent with the interpretation that different regions of the protein are denaturated to varying degrees at a specific GdnHCl concentration, such as in the case of carbonic anhydrase B (Wong andTanford, 1973). The conclusion that pyruvate kinase in the absence of ligands undergoes a multistep denaturation is consistent with the elegant denaturation studyon the enzyme by Doster and Hess (1981). These regions of differential stability are probably related to thedomains observed in cat muscle pyruvate kinase by x-ray crystallography (Stammersand Muirhead, 1977; Stuart et at., 1979). The effect of Phe on pyruvate kinase denaturation is significant. The denaturation process monitored by change in absorbance at 288 nm shows a more sigmoidal relationship (Fig. 4) in the presence of Phe. The increase in sigmoidicity implies a higher degree of cooperativity in the denaturation process. This conclusion is in agreement with the results by monitoring the changes in absorbance between 300 and 288 nm (Fig. 3), which could be interpreted to reflect a region or domain that is more susceptible to GdnHCl denaturation in the absence of Phe. The rest of the protein denatures only upon exposure to higher concentration of denaturant. In the presence of Phe, this region of pyruvate kinase is apparently more stabilized, and it assumes a stabilitythat is more typical of or its unfolding is more coupled to therest of the molecule. Hence, it is to be expected that thedenaturation of the whole molecule of pyruvate kinase would become morecooperative. The spectroscopic data are in good agreement with the hydrodynamic results, which also help to identify the nature of stabilization. Both the sedimentation and viscosity data
indicate clearly that the presence of Phe stabilizes the pyruvate kinase quaternary structure from GdnHCl denaturation. At 1.1M GdnHCl, the enzyme is dissociated into its subunits, and they are nearly completely unfolded. However, in the presence of 15 mM Phe, the enzyme maintains its tetrameric structure, butis probably partially unfolded. Thus, binding of Phe probably not only stabilizes a region of pyruvate kinase from denaturation at low concentration of denaturant, but also stabilizes the quaternary structure of the enzyme. Having established that the binding of Phe to pyruvate kinase most likely induces changes in domain-domain interactions, it is of interest to identify thedomain(s) which undergoes the most significant change. The most probable candidates are the domains that contain tryptophan residues. Let us review the experimental evidences that lead to this conclusion. Depending on the wavelength employed, the spectroscopic techniques monitor predominantly the behavior of either tryptophan or tyrosine residues in pyruvate kinase. Measuring denaturation using absorbance change at 288 nm probably reflects mainly the environmental changes around tyrosine residues, since there are 9 tyrosine residues as compared to 3 for tryptophan. The transition curves are monophasic. These results imply that, upon denaturation, the tyrosine residues in pyruvate kinase wouldprovide a signal more reflective of the whole molecule. This interpretation is consistent with the location of these tyrosine residues. The 9 tyrosine residues of cat muscle pyruvate kinase are distributed evenly throughout each subunit, 3 in each of the three domains. Therefore, this chromophore is probably an indicator of global structure. In contrast,absorbance change at 300 nm and fluorescence excitation at 295 nm yield results more reflective of the environment of tryptophan residues. The presence of Phe affects the tryptophan environments most significantly (e.g. almost a complete elimination of the increase in fluorescence intensity at low GdnHCl concentration, as shown in Fig. 1; and conversion of a biphasic denaturation transition to that of a monophasic, as shown in Fig. 3). In muscle pyruvate kinase, there are only 3 tryptophan residues in each subunit, and they are distributed as 2 residues in the C domain and 1in the B domain (Muirhead et al., 1986; Stuart et al., 1979). These domains are separatedfrom each other by the A domain of about 25-30 X lo3 inmass. Hence, the spectroscopic evidence implies that, upon formation of the pyruvate kinase-Phe complex, the tryptophan residue(s) in the B or C domain or both is perturbed. In addition to spectroscopic results, chemical evidence presented suggests that the hinge region between the A and B domains is perturbed by the presence of ligands. Results from partial tryptic digests of pyruvate kinase indicated that this method may be a relatively sensitive probe for enzyme conformation in solution. Under all conditions studied, tryptic digest patterns showed the same set of peptides being produced, as judged by their apparent molecular weight onSDSpolyacrylamide gel electrophoresis. However, there was a change in the rates of peptide production under some conditions. This can be interpretedas that pyruvate kinase in solution exists as anequilibrium between two conformations. The presence of Phe shifts the equilibrium toward one of the states, resulting in an increased exposure of the digestion sites. Likewise, Mg2' shifts the equilibrium toward the other state, leading to a population of pyruvate kinase molecules in which the tryptic digestion sites are less accessible. In summary, the evidence presented herein is consistent with domain movement in pyruvate kinase in response to ligands. The nature of this movement is more fully explored
Structure Solution by small angle neutron scattering and molecular modeling described in the following paper (Consler et al., 1988). REFERENCES Ainsworth, S., and MacFarlane, N. (1973) Biochem. J. 1 3 1 , 223-236 Boyer, P. D. (1962) in The Enzymes (Boyer, P. D., Lardy, H., and Myrback, K., eds) Vol. 6, p. 95, Academic Press, New York Carminatti, H., Jimenez DeAsua, L., Leiderman, B., and Rozengurt, E. (1971) J. Biol. Chem. 246, 7284-7288 Consler, T. G., Uberbacher, E. C., Bunick, G . J., Liebman, M. N., and Lee, J. C. (1988) J. BWL. Chem. 263,2794-2801 Dann, L. G., and Britton, H. G . (1978) Biochem. J. 1 6 9 , 39-54 Doster, W., and Hess, B. (1981) Biochemistry 2 0 , 772-780 Field, D. J,, Collins, R. A., and Lee, J. C. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,4041-4045 Ginsburg, A., and Carroll, W. R. (1965) Biochemistry 4, 2159-2174 Grant, G . A., Sacchettini, J. C., and Welggus, H. G. (1983) Biochemistry 2 2 , 354-358 Hassett, A., Blattler, W., and Knowles, J. R. (1982) Biochemistry 21, 6335-6340 Kayne, F. J., and Price, N.C. (1972) Biochemistry 11,4415-4420 Kayne, F. J., and Price, N.C. (1973) Arch. Biochem. Biophys. 159, 292-296 Kwan, C.-Y., and Davis, R. C. (1980) CUR.J. Biochem. 58, 188-193 Kwan, C.-Y., and Davis, R. C. (1981) Can. J. Biochem. 59,92-99
of Pyruvate Kinase
2793
Lee, J. C., Gekko, K., and Timasheff, S. N. (1979) Methods Enzymol. 6 1.26-49 Lonberg, N., and Gilbert, W. (1983) Proc. Natl. Acad. Sci. U. S. A. 80,3661-3665 Mildvan, A. S., and Cohn, M. (1965)J. BWZ. Chem. 240, 238-246 Mildvan, A. S., and Cohn, M. (1966) J. BWl. Chem. 241, 1178-1193 Monod, J., Wyman, J., and Changeux, J . P. (1965) J. Mol. Biol. 12, 88-118 Muirhead, H., Clayden, D. A., Barford, D., Lorimer, C. G . , FothergillGilmore, L.A., Schiltz, E., and Schmitt, W. (1986) EMBO J. 5, 474-481 Oberfelder, R. W. (1982) Ph.D. thesis, St. Louis University Oberfelder, R. W., Lee, L. L.-Y., and Lee, J. C. (1984a) Biochemistry 23,3813-3821 Oberfelder, R.W., Barisas, B. G., and Lee, J. C. (198413)Biochemistry 23,3822-3826 Pace, N. C. (1975) CRC Crit. Reu. Biochem. 3, 1-43 Privalov, P. L. (1979) Adu. Protein Chem. 3 3 , 167-241 Stammers, D. K., and Muirhead, H. (1975) J. Mol. Biol. 95, 213-225 Stammers, D. K., and Muirhead, H. (1977) J . Mol. Biol. 112, 309316 Stuart, D. I., Levine, M., Muirhead, H., and Stammers, D. K. (1979) J . Mol. Biol. 134, 109-142 Tanford, C. (1955) J . Phys. Chem. 59, 798-799 Wong, K. P., and Tanford, C. (1973) J. Biol. Chem. 248,8518-8522 Yphantis, D. A. (1964) Biochemistry 3 , 297-317
