Characterization of Slowly Interconvertible States of Phosphoribosyladenosine Triphosphate Synthetase Dependent on Temperature, Substrates, and Histidine*

Abstract Phosphoribosyladenosine triphosphate (PR-ATP) synthetase from Salmonella typhimurium was found to undergo a slow temperature- and ligand-dependent activation. The kinetics of activation appeared first order with half-lives up to 17 min depending on the conditions. Activation by the substrate ATP diminished the extent of subsequent temperature activation. The extent of activation appeared to be independent of pH and buffer concentration over limited ranges and was reversible. Two high temperature (25–37°) states of PR-ATP synthetase having different inhibition responses toward histidine exist depending upon prior ligand action. Apparent patterns of histidine binding determined from kinetic data are similar to those determined from equilibrium data obtained by dialysis and protein fluorescence enhancement. The possible metabolic importance of the high temperature "hysteretic" behavior is discussed.

was found to undergo a slow temperature-and ligand-dependent activation. The kinetics of activation appeared first order with half-lives up to 17 min depending on the conditions. Activation by the substrate ATP diminished the extent of subsequent temperature activation.
The extent of activation appeared to be independent of pH and buffer concentration over limited ranges and was reversible.
Two high temperature (25-37") states of PR-ATP synthetase having different inhibition responses toward histidine exist depending upon prior ligand action. Apparent patterns of histidine binding determined from kinetic data are similar to those determined from equilibrium data obtained by dialysis and protein fluorescence enhancement. The possible metabolic importance of the high temperature "hysteretic" behavior is discussed. 1 The abbreviations used are : PK-ATP, A-1-(5'.phosphoribosyl)adenosine triphosphate; PIIPP, 5-phospho-a-n-ribosylpyrophosphate; DTT, dithiothreitol; Hepes, A'-2-hydroxyethylpiperazine-A"-2-ethanesulfonic acid.
(4) have isolated a phosphoribosyl covalent enzyme intermediate, thus proving a double displacement mechanism which could operate in the absence of a ternary complex. Whitfield (2) found histidine inhibition to be cooperative but incomplete while Martin (1) and Bell and Koshland (3) found it to be uncooperative and complete The recent studies described in the accompanying manuscripts indicate two factors which might help explain these discrepancies. In the first place the association-dissociation behavior is complex (5). Although nominally a hexamer of identical subunits, PR-ATI' synthetase can be dissociated to a dimer at pH 10 and low ionic strength.
Also, low temperature or high ionic strength leads to a more comples aggregation.
In the second place, PR-ATP synthetase was found to contain a histidase impurity (7). Since histidase produces urocanic acid which absorbs in the ultraviolet spectral region, appreciable errors in the kinetic assays of PR-ATP synthetase occurred when high levels of histidine were present.
The production of a PR-ATP synthetase free of histidase activity and the understanding of the complex association-dissociation behavior of the enzyme has led us to re-examine some kinetic and equilibrium behavior of PR-ATP synthetase. During these studies a temperature-and ligand-dependent activation of the enzyme has been observed.
A coherent picture emerges which appears to explain the previous inconsistencies.

EXPERIMESTAL PROCEDURE
Male&&-Chromatographically pure Pl1-ATP was a gift from B. N. Ames, University of California. All other chemicals were obtained from the sources previously listed (7). Pure PR-ATP synthetase was prepared from the strain TA2165 by our new method (7) or from the histidine Eli or 01S4Z mutants (8) as noted. TA2165 enzyme was stored in liquid nitrogen (7) whereas Elf enzyme was stored in 50yo glycerol at -50" (3); 0184s was used immediately.
Melhods-TA2165 enzyme was transferred in all cases into the incubation buffers by Sephadex G-50 gel filtration at room temperature while Efl was done bv an 8-fold dilution from storage. The standard buffer is composkd of 0.10 M NaCl, 0.01 M Trisy 0.5 mM EDTA. and 1 mM DTT adiusted to either DH 7.5 with Heoes or pH 8.5'with HCl.
Unless otherwise noted,all pH measure'&ents were taken at room temperature.
Enzyme was incubated before assay for at least 90 min when at 0" and for 30 min when at 37" in those cases where equilibrium was desired. Slandard Assay-The standard assay is based on that of Voll For most histidine inhibition studies, a small aliquot of concentrated histidine which had been adjusted to pH 8.5 was also added. Enzyme was added to start the reaction at 25" which was monitored by the absorbance increase at 290 nm due to the formation of PR-ATP (10). Volumes of enzyme smaller than 1 ~1 were pipetted with a 5-~1 microsyringe (Hamilton Co., Whittier, California).
Our unit of activity is defined as a change in absorbance of 0.10 per 5 min and corresponds to 1.75 nmoles of product per min based on an extinction coefficient of 3.6 X lo3 for PR-ATP (11 The titration also was conducted in the presence of 0.5 mM PRPP(NaJ, 10 mu MgCl*, and 1.5 units of yeast inorganic pyrophosphatase after prior incubation with the enzyme for 1 min. A titration in the presence of 10 rnM ATP and 20 mM MgClz was possible by using 285-nm excitation and a somewhat bigger concentration of enzyme. A similar titration was performed in the last case at pH 8.5 also.

Standard Assay after Some Lliferent Preincubation
Conditions Several different shapes of product versus time curves were obtained when identical concentrations of PR-ATP synthetasc which had been incubated under different conditions were assayed. Fig. 1 shows some typical assay curves. Linear assays were obtained for enzyme in the prior presence of histidine or at 37", but not at 0" in the prior absence of histidine.
The lag phase demonstrated that the enzyme gained activity for several minutes following addition to the assay buffer after which a linear product versus time curve results. The linear limiting activity of the 0" zero histidine enzyme, which was lower than that resulting from the other preincubation conditions, was measured in the activity plots to be discussed below. Reversible Temperature-dependent A&i&ion-The rate of the transformation of the 0" form to a higher temperature form of the enzyme is shown in Fig. 2 2. Temperature-dependent activation of PR-ATP synthetase.
El1 PR-ATP synthetase in 0.10 M Tris-HCl, 0.15 M KCl, and 10 mM MgClt at pH 8.5 was incubated at 0" for 2 hours before the temperature was raised to 22" at time zero by agitating gently in a 22' water bath. Aliquots (10 ~1) were removed and assayed in the standard assay at 25" at the times indicated. After 75 min, the remaining solution was placed in an ice bath and aliquots were removed and assayed at 25" at the times indicated.
a solution of the enzyme in a small test tube was raised from 0 to 22". A slow activation of PR-ATP synthetase occurred. When after 75 min no further increase in activity was observed, the same enzyme solution was transferred back to an ice bath.
Aliquots were taken as before for assay at 25". A slow loss of activity occurred until the activity reached its initial (time zero) level. PR-ATP synthetase undergoes a slow, apparently reversible, temperature-dependent activation. The kinetics of activation from the "less active" 0" form of the enzyme to a "more active" 22" form of the enzyme was first order as was the kinetics of the reverse process (semilog graphs not shown).
The half-life for the 0 ---t 22" transition was 8.9 min, while that for the 22 -+ 0" transit'ion was 17.6 min. E$ect of Temperature on Extent and Rate of Activation,-The effect of temperature on the activation process is reported in Elf enzyme was prepared as described in Fig. 2. Each tube was held at 0" at least 2 hours prior to the start of the experiment when each tube containing 200 ~1 of the enzyme solution was transferred to a water bath at the temperature indicated.
A time zero point was taken for each experiment at 0' prior to placing the tube in the appropriate water bath.
of PR-ATP synthetase was dependent upon the incubation temperature.
The rate of the activation process was also dependent upon the temperature. The rates obtained from semilog p1ot.s were plotted in an Arrhenius plot (not shown). The activation energy for the temperature-dependent activation was 14.7 f 1.3 Cal per mole. At 25", a AFI of 19.6 Cal per mole, a AH# of 14.1 f 1.3 Cal per mole, and a ASf of -18.5 f 4.4 cal per degree mole were calculated. E$ect of pH on Temperature-dependent Activation--The effect of pH on the temperature activation of PR-ATP synthetase was investigated. There was essentially no pH effect on the activation process over the pH range 7.5 to 8.4 measured at 25". This eliminated a pH change as the source of the activation process (the pH for this Tris buffer system decreased nearly 0.5 pH units when the temperature was raised from 0 to 25"). The slow activation also occurred in N, N'-bis-2-hydroxyethyl glycine buffer at pH 8.5 in the same manner as in the Tris buffer.

Conjormationd
Probes--The optical rotation of a solution of PR-ATP synthetase was measured at 233 nm as the solution warmed from 0 to 25". No changes were detected. No detectable differences in ultraviolet spectra or fluorescence emission spectra were observed between the "activated" and "nonactivated" forms of PR-ATP synthetase. Probably no major changes in subunit structure occur upon activation, but rather minor shifts of important residues which affect gross spectroscopic parameters only slightly.
Activation of PR-ATP Synthetase at 0" by Substrates-Zerodegree enzyme was found to be activated by subst.rate ligands also. The effect of ATP is shown in Fig. 4. The addition of 10 mM ATP to the enzyme at 0" resulted in a a-fold activation over a 90-min time period. The kinetics of the ATP-dependent activation appeared first order and had a half-life of 19 min.
The activation of PR-ATP synthetase at 0" by ATP was concentration dependent ( Table I). The ATP concentration for half-maximal activation at 0" was about 500 pM which is similar to the ~0.~ of the enzyme for ATP (2, 3). When PR-ATP synthetase was incubated with other substrate ligands at O", an activation also was observed ( Table I).
Activation of PR-ATP Synthetase at 0" by Hi&&e--The histidine-dependent activation of 0" enzyme seen in Fig. 1 was investigated under equilibrium conditions as shown in Fig. 5. The specific activity of the fully histidine-activated enzyme was about 2.5 times that of the unactivated enzyme. The activation process was cooperative with a Hill constant of 2.4 and the midpoint occurred at 0.071 mM histidine. Assays became increasingly more linear at zero time after addition of enzyme as the histidine incubation concentration was increased. Dependence of Activation on Protein Concentration-The histidine activation factor for the 0" enzyme was variable in different experiments, depending upon the enzyme concentration. The possibility of an enzyme dissociation could be checked by carrying out a dilution study in the presence and absence of incubation histidine by assaying larger aliquots of the more dilute enzyme. There was no effect on the apparent specific activities of 0" enzyme arising from varying the incubation concentration over the range 6 to 0.09 mg per ml of enzyme. The histidine activation factor was about 1.8 in this experiment. The variable histidine activation factor was not related to the enzyme concentration during preincubation.
Another possible concentration effect might arise from the amount of enzyme actually in the assay solution. Fig. 6 shows an experiment in which increasing aliquots of a single enzyme solution incubated at 0" in zero histidine standard buffer were  5. Apparent activity of 0" enzyme in presence of histidine. The increase in apparent activity in t,he standard assay for TAB65 enzyme incubated at 0' at pH 7.5 is plotted as a function of histidine in the incubation solut,ion. Inset is a Hill plot of t,he data. K, is the activation midpoint while 7~ is the cooperativit,y constant.
Similar results were obtained for enzyme incubated at pH 8.5. assayed.
The specific activity decreased as the enzyme corlcentration in the assay solution increased. A similar experiment on histidine-preactivated enzyme showed no such assay dilution effect.
The decrease in activity of the zero histidine enzyme might be due to a transformation of the enzyme from an active dissociated state at low assay concentrations to an inactive aggregated state at high assay concentrations. This possibility was checked by analyzing the shape of the supposed dissociation curve as discussed in the legend to Fig. 6. These calculations show that all reasonable permutations for a dissociation-association explanation are discarded. Furthermore, ~~0,~ for 0" zero histidine enzyme in the assay solution previously was determined (5) by the Cohen method at a layered initial concentration about in the middle of the dilution activation curve (Fig. 6). The enzyme sedimented as a well-behaved hexamer.
There was no indication of a second enzyme species, or of any dissociation or aggregation reaction.
The apparent, activity in the standard assay of 0" zero-histidine enzyme from TA2165 at pH 7.5 is plotted as a function of the concentration of enzyme in the assay cuvette. Apparent specific activity drops at high enzyme concentrations. The ilxet shows log-log plots of normalized second and third order dissociations (6).
In this method, the amount of dissociated species, say D,, is plotted vwsus total protein divided by L), for second and third order dissociations labeled d and 3, respectively. A first order reference labeled 1 is included. We need to make plausible assumptions about the activities of the fully dissociated and associated enzyme states to analyze the shape of the experimental curve.
Assuming that the fully dissociated state has the activity of the histidine-activated enzyme (1950 units per ml here) and that the fully aggregated state has no activity we obt,ained the heavy line in the insel. The data are far from fitting any equilibrium enzyme dissociation reaction using the above assumption about extreme states.
In fact, all other reasonable assumptions, such as that the fully dissociated stat,e gives 1120 units per ml and the fully aggregated stale gives 380 units per ml, result in poor fits of any possible whole order dissociation. The Cohen active enzyme-substrate complex sedimentation velocity technique was used 10 study the aggregation state at the concentration indicated by the arrow (5). Enzyme behaved as a hexamer of 8.7 s~o.ul.
The most straightforward explanation for the apparent dilution activation is that it is caused by product inhibition. The onset of product inhibition under these conditions can be demonstrated by pre-adding PR-ATP to the standard assay. Fig. 7 presents the inhibition observed using 0" zero histidine enzyme. Significant product inhibition occurs when about 0.1 absorbance unit of PR-ATP has been formed. It is to be remembered that 0" zero histidine enzyme does not attain full activity for several minutes.
Thus, product inhibition sets in before all of the enzyme has been activated when a high concentration of enzyme is assayed.
The opposing activation-inhibition effects produce approximate linearity in the assay for a period of time. The enzyme appears to have a lower specific activity than an identical sample at lower concentration, but this is an artifact. The maximal specific activity which 0" zero histidine enzyme can exhibit under these assay conditions consequently is uncertain. The highest observed has been about 1700 units per ml per 280nm absorbance unit of enzyme. of the activation at 23" to that at 0" in the presence of increasing amounts of ATP was found to decrease almost to 1.0 at 5 mM ATP.
Thus, an increase in either temperature or ATP concentration seems to induce a similar activity state of the enzyme. Arrhenius Plot jor Temperature-dependent Forms of PR-A TP Synthetases-To investigate further the effect of temperature upon PR-ATP synthetase, initial rates were measured for 0 and 23" forms of the enzyme at different assay temperatures. An Arrhenius plot is shown in Fig. 9. The two curves are nearly parallel and both show a sharp break between 16 and 20". Calculation of the energy of activation for product formation from the 4 to 16" arm of the curve for the 0 and 23" forms revealed energies of activation of 43.0 and 42.2 Cal per mole, respectively. The energy of activation calculated in the 20 to 40" arm of the curve for the 0" enzyme form is 18.3 Cal per mole while a value of 13.9 Cal per mole was calculated for the 23" enzyme form. The similarity of the slopes and the breaks in the Arrhenius plots at nearly the same temperature suggest an over-all structural change in the protein which is common to both the activated and nonactivated forms of the enzyme. 7. Product inhibition of the 0" zero-histidine enzyme. TA2165 enzyme preincubated in standard buffer at pH 7.5 was assayed in the presence of increasing amounts of Pit-ATP preadded in the standard assay procedure.
The reciprocal of initial velocity is plotted versus concentration of PK-ATP. One-half inhibition occurred at 0.17 mM PK.ATP. This corresponds to an absorbance at 290 nm of 0.6. specific activity from 0" histidine-activated enzyme (cf. Fig. 1). However, a difference between the two 37" enzymes became apparent when histidine feedback inhibition was studied. Fig.  10 shows a plot of the residual activity of enzyme preincubated in the presence and absence of hist,idine as a function of the amount of histidine in the assay buffer. The histidine-preincubated enzyme was inhibited 100y6 by 0.2 mM assay histidine while the histidine naive enzyme retained 120/o activity even in the presence of 8.7 mM assay histidine.
This residual activity decreased to zero after several minutes in the high histidine assay solution, thus giving a concave down assay trace as shown by the dashed line in Fig. 1. In other words, 37" zero histidine-incubated enzyme cannot be totally inhibited instantaneously, but does relax within a few minutes under assay conditions to a form which is totally inhibited.
Three Enzyme States Dependent on Temperature and Ilistidine- Table II summarizes some kinetic behavior in the standard assay of PR-ATP synthetase incubated in standard buffer at pH 7.5 under the four conditions of Fig. 1, The histidine-preincubated enzyme is more cooperatively inhibited by histidine in the assay both at 0 and 37". The apparent specific activity for the 0" zero histidine enzyme is variable but clearly lower than for the other enzyme states.
On the basis of Table II there are three  discernible states of the enzyme, one each at 0 and 37" in the absence of histidine and the one resulting from histidine incubation at both temperatures. Very similar results were obtained from enzyme incubated at pH 8.5 under otherwise similar conditions. Elf enzyme was prepared as described in Fig. 2. ATP was added to the concentrations indicated and each solution was assayed after incubation for 3 hours at 0". Then each solution was transferred to 23" and assayed after a Y-hour incubating period. The ratio of the 23 to the 0" activities is plotted as a function of ATP concentration.

Arrhemus
Plot for a 0" and a 23" Form El1 enzyme was prepared as described in Fig. 2. Half of the enzyme solution was incubated at 0 f 0.3" for 2 hours prior to the start of the experiment, while the other half was incubated at, 23 Z+Z 0.3". Initial rates were measured in duplicate in 0.2 M Tris, lG.7 mM MgC12, 0.15 mM KCl, 5 mM ATP, and 0.5 mM PRPP.
Yeast inorganic pyrophosphatase was added to each assay. The pH of the assay buffer was adjusted to 8.5 at each temperature.
A plot of the logarithm of the velocity versus reciprocal absolute temperature is shown.
above were obtained with the "standard assay," which involved addition of enzyme to initiate the reaction. However is plotted as a function of assay histidine. Assays of the zero-histidine-incubated enzyme were concave down at high histidine concentrations as indicated by the dashed line in Fig. 1.

TAISLE II
The Hill constant (no), midpoint of inhibition by histidine in the assay (K;), specific activity per absorbance unit at 280 nm, shape of the assay trace, and per cent inhibition by high histidine are given.
These are the results for incubation in standard buffer at pH 7.5. Essentially the same results were obtained for incubation at pH 8.5. The inverse lag under 37" and no histidine occurs in the presence of high histidine in the assay. the order of addition of substrate and enzyme to the assay solution.
This was found to be the case as shown in Fig. 11. The histidine feedback inhibition patterns for similar solutions of enzyme were obtained in one case by initiating the reaction with PRPP (ATP incubation) and in the other case by initiating it with ATP (PRPP incubation).
Varying levels of histidine were in the assay buffer and enzyme was added next to last. Enzyme preincubated in the presence of ATP could not be totally inhibited by high levels of histidine and the inhibition response was noncooperative.
On the other hand, PRPP-incubated enzyme was essentially totally inhibitable in a highly cooperative manner. Thus there are at least two substrate-dependent enzyme states possible.
Histidine The binding is not comparable to that in Fig. 13 for reasons discussed in the text.
The histidine saturation level corresponded to 5.2 g moles of histidine per 216,000 g of enzyme, assuming an enzyme specific activity (standard assay on histidine-activated enzyme) of 2400 units per mg of enzyme (7). This amount of histidine is slightly below the expected 6, and may be due to enzyme denaturation during 2 days of dialysis. It is unlikely to be caused by contamination by histidase in this preparation of enzyme (7) since a saturation plateau is apparent in Fig. 12. Fig. 13 illustrates a series of histidine equilibrium binding studies performed by enzyme fluorescence enhancement titration at 25" and pH 7.5. The top frame is the histidine binding curve obtained in the absence of other ligands. Half saturation occurred at 0.29 mM with a Hill constant of 2.1. In the presence of 0.5 mM PRPP, 10 mM MgC&, and inorganic pyrophosphatase, the histidine binding became stronger and more cooperative with a Ki of 0.19 mM and a Hill constant of 2.9 being obtained. The increase in fluorescence of the TA2165 enzyme in the presence of histidine in standard buffer at pH 7.5 and 25" is shown. The bottom two curves were obtained in the presence of the indicated substrate ligands.
The Hill parameters were determined on separate plots and are also given. The 0" zerohistidine form of enzyme prepared by the older method was utilized.
the presence of 10 mM ATP and 20 rnx MgCl, the binding became weaker and less cooperative, with a Ki of 0.37 rnM and a Hill constant of 1.9 being obtained.
At pH 8.5 in the presence of MgATP the histidine dissociation constant increased to 0.82 mM while the Hill constant decreased to 1.7. Thus histidine binding is perturbed by the substrates.

Temperature-and Time-dependent Conformation
Change-The enzyme PR-ATP synthetase undergoes a temperature-dependent activation process which is apparently caused by an equilibrium between conformational states of the protein which interconvert at rates slower t,han the time span of a typical 5-min assay. The temperature-dependent process is reversible and can be influenced by substrates and the allosteric effector histidine.
Two different 37" forms of the enzyme which depend upon histidine were detected by kinetic analysis. Roth of the 37" enzymes were shown in a previous paper (5) to be hexamers.
Thus, the kinetic differences between the two 37" enzymes apparently lie in different slowly interconvertible conformational states of the hesamer which regulate both the cooperativity of histidine inhibition and residual activity at high histidine levels. This residual activity is not due to a partial desensitization toward histidine as a result of "aging" as has been observed before (l), since the enzyme does become inhibited after a few minutes. There appears to be no difference on the basis of the kinetic studies presented here or the previously mentioned physical studies between histidine-liganded enzyme at 0 or 37". Histidine appears to stabilize one particular hexameric state of the enzyme against temperature changes.
Enzyme purified in two different ways from the different sources utilized here responds in the same manner insofar as the temperature activation.
There is no reason to suspect that all the enzyme preparations are not essentially identical in PR-ATP synthetase properties.
The conversion from the low to a high temperature form of the enzyme occurs primarily above 11" under incubation conditions and appears to be essentially complete at 37". The physical basis for the difference between 0" and 22-37" enzyme was elucidated in an accompanying paper (5). Low temperature enzyme (4-8") exists in a continuous distribution of aggregated species dependent upon protein concentration while high temperature enzyme (22-37') exists predominately as a hexamer.
In addition to the enzyme states which depend upon incubation temperature and histidine conditions, there are stat,es which depend upon prior substrate action. Activation of 0" enzyme occurs in the presence of ATP, PRPP, and Pl'i.
This probably is the result of stabilization of hexameric enzyme, since both ATP and PRPP have been shown to produce hexamers from dimeric enzyme under other conditions (5). Incubation of the cnzymc at 25" and pH 8.5 in the presence of ATP abolishes cooperativity of histidine inhibition and induces a residual enzymatic activity at high histidine levels. This behavior is very similar to that resulting from incubation in zero histidine at 37", and may indicate that the two sets of conditions promote the same hexameric enzyme state.
Enzyme incubated in the presence of PRPP at 25" is inhibited totally in a very cooperative manner by histidine. This behavior is similar to that resulting from histidine incubation. Thus, there is compelling evidence only for three different slowly interconverting states on the basis of the kinetic evidence presented here. Different 0" forms of the enzyme which depend on the presence Equilibrium Binding of Histidirr-Slightly less than 6 g moles or absence of histidine were detected by kinetic analysis. The of histidine are bound at 4" at saturation by 216,000 g of the physical basis for these forms also has been elucidated elsewhere by the demonstration that 0" histidine-activated enzyme exists as the hexamer (5). Since it exhibits a pronounced lag upon assay, the 0" indefinitely aggregated enzyme is initially inactive or of greatly reduced activity. III addition, this enzyme never achieves an activity state during assay which exhibits kinetic parameters similar to those of the histidine-activated enzyme. Nevertheless, sedimentation velocity analysis of the two forms of the active enzyme-substrate complex (5) demonstrated that both forms of the enzyme are hexamers under assay conditions. Apparently, indefinitely aggregated low activity 0" zero histidine enzyme reassociates within several minutes in the assay buffer to give an active hexamer, this process being the cause of the kinetic lag. An apparent dilution activation of 0" zero histidine-incubated enzyme was shown to be an artifact caused by product inhibition and the time lag. Since both the cooperativity of histidine inhibition and the specific activity are significantly decreased, it is as if the rapidly assembled histidine-naive hexamer does not contain properly "interlocked" subunits. Dilution activation of Escherichia coli PR-ATP synthetase has been reported by Kryvi and Klungsoyr (14). The E. coli enzyme may be different in that no assay lag has been reported, and the dilution activation occurred after incubation in an assay solution containing ATP, which stabilizes the hexamer form of the Salmonella enzyme (5). Therefore, it cannot with certainty be concluded that the dilution activation phenomena of the two enzymes are the same.