Chorismate Mutase-Prephenate Dehydratase

The bifunctional enzyme chorismate mutase-prephenate dehydratase catalyzes the first two reactions specific for phenylalanine biosynthesis in Salmonella fyphimurium. Both activities are subject to feedback inhibition by phenylalanine. In the presence of phenylalanine the sedimentation coefficient of the enzyme is increased from approximately 5.6 to 8.0 S and the molecular weight is increased from approximately 109,000 to 220,000. This apparent dimerization is dependent upon the concentrations of phenylalanine and enzyme. Chorismate mutase-prephenate dehydratase that has been rendered insensitive to feedback inhibition either chemically or by mutational alteration does not dimerize. Therefore, dimerization appears somehow related to end product control. However, dimerization is not obligatory for feedback inhibition of either activity. At low protein concentration some enzyme does not dimerize in the presence of phenylalanine. Such enzyme can still be inhibited by phenylalanine. It is therefore suggested that phenylalanine binds to the monomer (smallest active species) to produce an inhibited monomer which dimerizes at appropriate protein concentrations.

In the presence of phenylalanine the sedimentation coefficient of the enzyme is increased from approximately 5.6 to 8.0 S and the molecular weight is increased from approximately 109,000 to 220,000. This apparent dimerization is dependent upon the concentrations of phenylalanine and enzyme.
Chorismate mutase-prephenate dehydratase that has been rendered insensitive to feedback inhibition either chemically or by mutational alteration does not dimerize. Therefore, dimerization appears somehow related to end product control.
However, dimerization is not obligatory for feedback inhibition of either activity. At low protein concentration some enzyme does not dimerize in the presence of phenylalanine.
Such enzyme can still be inhibited by phenylalanine.
It is therefore suggested that phenylalanine binds to the monomer (smallest active species) to produce an inhibited monomer which dimerizes at appropriate protein concentrations.
The first two reactions specific to phenylalanine synthesis in microorganisms are catalyzed by chorismate mutase (Equation 1) and prephenate dehydratase (Equation 2).
Chorismate -+ prephenate (1) Prephenate + phenylpyruvate + CO2 + H?O (2) These activities are physically associated with a bifunctional enzyme or enzyme complex in Eschekhia co&, Aerobacter aerogenes, and Salmonella typhimurium (1,2). The native enzyme is designated chorismate mutase-prephenate dehydratase. Both activities from S. typhimurium are subject to feedback inhibition by phenylalanine and cooperative kinetics is observed (2). The results of previous experiments have suggested that the two catalytic sites are functionally distinct and that the regulatory site(s) is distinct from the catalytic sites (3).
We previously reported that phenylalanine caused an increase in the sedimentation coefficient of chorismate mutase-prephenate dehydratase (2). X4any papers have appeared in the literature concerning association-dissociation reactions of regulatory enzymes (5-28).
We report here experiments which show that chorismate mutase-prephenate dehydratase undergoes a phenylalanine-dependent dimerization. This dimerization is related to feedback inhibition by phenylalanine but appears not to be obligatory for inhibition.
A preliminary account of these results has been reported (4).

Materials
Sub&ales-Barium chorismate and chorismic acid were isolated from A. crerogenes strain 62-l (29, 30) and were converted to the potassium salt before use. The concentration of potassium chorismate solutions was determined by enzymatic assay (31). Barium prephenate was obtained by the method of Gibson (29) and was converted to the potassium salt with a 1.5-fold excess of KzS04. Prephenate was assayed by acid-catalyzed conversion to phenylpyruvate (2). Phenylalanine-substituted Sepkarose-Phenylalanine-substituted Sepharose was prepared according to the general method of Cuatrecasas, TYilchek, and Anfinsen (32) as previously described (3).
Bacterial Strains-The strains of S. typhimurium, derived from wild type strain LT2, were aro A48, aro A48 tyr 25, and SA 34. The first two strains were used as sources for native chorismate mutase-prephenate dehydratase while the last was the source for feedback-insensitive enzyme. These strains have been previously described (3).
Issue of October 10, 1971 J. C. Xchmit and H. Zallcin 6003 centration of phenylpyruvate from its absorption at 320 nm in 1 N NaOI-I.
One unit of activity corresponds to the conversion of 1 pmole of substrate to product in 10 min at 37". Specific activity is defined as units per mg of protein.
Protein was determined by the method of Lowry et al. as described by Layne (33) with bovine serum albumin as a standard or from absorbance measurements at 280 nm and 260 urn (33).
Growth of Bacteria-Strains (cro A48 or aro A48 tyr 25 were grown in 1 liter of media in 2-liter flasks as described previously (3). Alternatively, these strains were grown in 5-liter batches in a 7-liter New Brunswick FS307 fermentor by the method of Chen and Segel (34). The media contained, per liter: 12 g of glucose, 7 g of ammonium chloride, 1 g of sodium sulfate, 70.5 g of dibasic pot'assium phosphate, 12.9 g of monobasic potassium phosphate, 2 ml of trace elements (34), 175 mg of L-tryptophan, 400 mg of Ltyrosine, and 15 mg of L-phenylalanine.
The inoculum for each fermentor consisted of 80 ml of a culture grown overnight in nutrient broth.
Several drops of silcone antifoam were added and the culture was stirred vigorously with high air flow. The pH was adjusted to 7.0 with 10 N NaOH at l-hour intervals after the first 8 hours of incubation at 37". About 30 ml of 10 N NaOI-I were added during the 15.hour incubation period.
The cells were harvested in stationary phase after reaching a turbidity of about 3000 Klett u&s measured with a Klctt calorimeter using a No. 42 filter. The yield was approximately 18 g of cell paste per liter. The cells were washed with 0.05 M potassium phosphate buffer, pH 7.4, and stored at -10" to -20".
The specific activity for prephenate dehydratase of cells grown in the fermentor was from 0.8 to 1.3 units per mg.
Strain SA 32 (the source of phenylalanine-insensitive chorismate mutase-prephenate dehydratase) was grown in a-liter flasks as described for strain aro A48 (3).
Partial Purification of Chorismate Nutase-Prephenate Dehydratase-Partial purification to a specific activity of 28 to 66 for chorismate mutase and prephenate dehydratase activities was performed as described previously (2). Alternatively, enzyme of specific activity 130 to 250 was obtained by chromatography of a crude extract of either strain aro A48 or (lro A48 tyr 25 on a column of phenylalanine-substituted Sepharose. A typical purification procedure using affinity chromatography was as follows. Crude extract (100 ml, 15 mg of protein per ml) prepared by sonic disruption (2)  This was followed by 80 ml of buffer solution containing 10 mM potassium phosphate (PI-I 7.4), 0.4 rnhf dithiothreitol, and 0.2 rnhf EDTA. The enzyme was eluted from the column with buffer solution containing 2 m&I potassium phosphate (pII 7.4), 0.4 mM dithiothreitol, 0.2 mu EDTA, and 1 mM phenylalanine.
The peak fraction was dialyzed for 2 hours against 15 volumes of buffer solution containing 0.1 M potassium phosphate (pH 7.4), 0.4 m11 dithiothreitol, 0.2 mu EDTA, and 1 mM phenylalanine.
The yield based on chorismate mutase activity was approximately 25%. Feedback-insensitive chorismate mutase-prephenate dehydra-tase was partially purified from strain SA 34 as described previously (3). Sucrose Gradient Centrifugation-Linear 5 to 20% sucrose gradients (4.5 ml) were prepared at room temperature with buffer and other additions as indicated in the figure legends. An additional 0.5 ml of 2O$& sucrose solution was added to the bottom of tubes centrifuged at 60,000 rpm. Gradients were equilibrated for 3 to 6 hours at 2". After centrifugation the tubes were punctured and lo-drop fractions were collected and assayed for enzyme activity. The s20,W value and approximate molecular weight of chorismate mutase-prephenate dehydratase were estimated (35) by comparison with yeast alcohol dehydrogenase.

flfect of Phenylalanine on Sedimentation Velocity and Apparent Molecular Weight of Chorismate Mutase-Prephenate
Dehydratase-The sedimentation coefficient of chorismate mutase-prephenate dehydratase was estimated by sucrose gradient centrifugation. Typical data are shown in Fig. 1. In the absence of phenylalanine a sedimentation coefficient of 5.7 S was obtained for both chorismate mut,ase and prephenate dehydratase activities (Fig.  la).
The value for ~20,~ was increased to 7.7 S when centrifugation was conducted in the presence of 1 mM phenylalanine ( Fig.  1B). An intermediate sedimentation coefficient of 6.6 S was obtained for centrifugation in 1 PM phenylalanine as shown in Fig.  1C. Under all conditions of centrifugation chorismate mutase and prephenate dehydratase activity profiles were always superimposable.
Both the slower and faster sedimenting forms of the enzyme were subject to inhibition by phenylalanine.
Phenylalanine caused an increase in the apparent molecular weight of the bifunctional enzyme as determined by gel filtration with Sephadex G-200 (Fig. 2). Gel filtration indicated that in the absence of phenylalanine ( Fig. 2A) the enzyme was smaller than either yeast alcohol dehydrogenase or lactate dehydrogenase and had a molecular weight of approximately 118,000. In the presence of 1 m&I phenylalanine, the elution position of the enzyme indicated a molecular weight of approximately 220,000, which is larger than either lactate dehydrogenase or alcohol dehydrogenase.
To each 4.5-ml gradient was added 0.1 ml of enzyme solution containing about 1.0 unit of chorismate mutase activity (30 pg of chorismate mutaseprephenate dehydratase) and 60 rg of yeast alcohol dehydrogenase. The enzyme solution had been dialyzed against buffer solution containing the same concentration of phenylalanine as was included in the gradient.
Centrifugation was at 35,000 rpm for 14 hours at 2' in the SW 39 rotor with a Spinco L2-65 B centrifuge. Fractions of 10 drops were collected.
The positions of the peaks for yeast alcohol dehydrogenase (~20,~ = 7.6 S) and chorismate mutase-prephenate dehydratase are indicated by UT~OWS. In B the position of yeast alcohol dehydrogenase was omitted from the figure because of its close proximity to the peak of chorismate mutase-prephenate dehydratase.
PHE, phenylalanine. FIG. 2 (right). The effect of phenylalanine on the apparent molecular weight of chorismate mutase-prephenate dehydratase as determined by gel filtration.
A column, 1.4 X 35 cm, of Sephadex G-200 was equilibrated with 200 ml of 0.1 M potassium phosphate (pH 7.4), 0.8 mM dithiothreitol, and 0.2 mM EDTA either in the presence (bottom) or absence (top) of 1 mM phenylalanine.
The same column was used for both determinations.
Chorismate mutase-prephenate dehydratase (0.75 ml containing 20 units of chorismate mutase activity) was dialyzed for 3 hours against the same buffer solution that was used to equilibrate the column. To this was added 0.15 mg of bovine liver catalase, 0.20 mg of yeast alcohol dehydrogenase, 0.10 mg of rabbit muscle lactate dehydrogenase, 10 mg of bovine serum albumin, and 1 mg of horseradish peroxidase. Fractions containing 10 drops (-1 ml) were collected at 5". Enzyme assays were conducted as described under "Methods." The position of the chorismate mutate activity is indicated by the dashed lines.
ADH, yeast alcohol dehydrogenase; LDH, lactate dehydrogenase; BSA, bovine serum albumin. mined by gel filtration and by sucrose gradient centrifugation is given in Table  Molecular weight was estimated from ~20,~ as described (35). Molecular weight was also estimated by gel filtration with Sephadex G-200. All reagents were tested at a concentration of 1 mM. At least 1 unit of chorismate mutase activity was used for each sucrose gradient centrifugation experiment. The conditions for gel filtration are given in Fig. 2 concentration on the sedimentation coefficient and molecular weight of' chorismate mutase-prephenate dehydratase. The sedimentat.ion coefficient wa.s estimated by sucrose gradient centrifugation at the indicated concentrations of phenylalanine or tyrosine.
In cases in which multiple determinations were made their number is indicated in parentheses and the range is indicated by the vertical arrow.
At least 1 unit of chorismate mutaseprephenate dehydratase activity was used for each determination. Cent.rifugation was at 35,000 rpm for 14 hours at 2" in an SW 39 rotor or at 60,000 rpm for 8.5 hours at 2" in an SW 65 rotor with a Spinco L2-65 B centrifuge.
varying phenylalanine concentrations on the sedimentation coefficient and apparent molecular weight of chorismate mutaseprephenate dehydratase was determined by sucrose gradient centrifugation and is shown in Fig. 3. In the absence of phenylalanine, an average value of 5.6 S corresponding to an approximate molecular weight of 88,000 was obtained from eight experiments. Phenylalanine at a concentration of 1 PM caused a partial shift to 6.9 S (mol wt approximately 120,000). A single sym-Issue of October 10, 1971 J. C. Xchmit and H. Zalkin metrical peak of enzyme activity was obtained (Fig. lC), indicating a rapid association-dissociation (46) under these conditions. At 5 PM phenylalanine and higher, a "complete" shift to 8.0 S, molecular weight approximately 154,000, was observed. Tyrosine caused an increase in apparent molecular weight for the enzyme but was not as efficient as phenylalanine (Fig. 3). At a tyrosine concentration of 100 PM a sedimentation coefficient similar to that obtained with 1 PM phenylalanine was found. With 1 mM tyrosine or 2 PM phenylalanine similar molecular weights were obtained.
Thus, it is estimated that 100 to 500 times more tyrosine is required to evoke a sedimentation shift similar to that caused by phenylalanine.
Neither 1 InM tyrosine nor 1 mM tryptophan, alone or in combination, modified the activity of chorismate mutase-prephenate dehydratase or changed the inhibition caused by phenylalanine.
Tyrosine at a concentration of 5 m&f did not inhibit chorismate mutase activity when assayed at 0.08 to 0.5 mM substrate.
Since the minimal concentration of phenylalanine required to attain 20% inhibition of chorismate mutase-prephenate dehydratase under these conditions was 40 pM, this level of inhibition by tyrosine might require a concentration of 4 to 20 my. back of inhibition by 5 mM tyrosine indicates that either inhibition and association are not necessarily related or that higher tyrosine concentrations are needed to obtain inhibition of enzyme activity. The insolubility of tyrosine precluded testing its effect at higher concentrations.
nn-P-Thienylalanine, an analogue of phenylalanine which inhibits both chorismate mutase and prephenate dehydratase activities, also caused aggregation (Table I). Tryptophan, chorismate, dithiothreitol, and phenylpyruvate, the product of prephenate dehydratase, did not affect the sedimentation velocity. The earlier report of aggregation in the presence of dithiothreitol (2) was an artifact caused by carryover of phenylalanine during preparation of the gradients. E$ect of Enzyme Concentration on Monomer-Dimer Equilibtium-Dimerization was dependent on enzyme concentration. At an enzyme concentration of 2.2 units per gradient more than 90% of the chorismate mutase ( Fig. 4A) and prephenate dehydratase ( Fig. 4B) activities sedimented as dimer (7.8 S) in the presence of 5 PM phenylalanine.
At loo-fold lower enzyme concentration (0.02 unit per gradient) both activities sedimented in broad activity profiles extending from the dimer region to the position where monomer would be expected (approximately 5 to 6 S). Yeast alcohol dehydrogenase was not used as a standard at low enzyme concentrations because it was contaminated with a trace of prephenate dehydratase activity.
The data in Fig. 4  (c) D esensitization of chorismate mutase activity was obtained at pH less than 6. The ability of the desensitized enzyme to undergo dimerization in the presence of phenylalanine was investigated in order to determine whether there is any correlation between feedback inhibition and dimerization.
It was shown previously that treatment of chorismate mutaseprephenate dehydratase with bromopyruvate resulted in loss of prephenate dehydratase activity (3). The remaining chorismate mutase activity was completely insensitive to inhibition by phenylalanine.
The data in Fig. 5 show that a major fraction of bromopyruvate-desensitized chorismate mutase activity remained monomeric in the presence of 5 pM phenylalanine (.s~~,~ = 5.0 S; mol wt approximately 74,000). A lesser fraction of the chorismate mutase activity was observed to associate. In a control esperiment with untreated chorismate mutase-prephenate dehydratase, under identical conditions of centrifugation, a single peak of activity at 7.8 S (mol wt approximately 149,000) was obtained.
Chorismate mutase activity from the peak fraction of the nontreated enzyme gave 66% inhibition when assayed with phenylalanine.
Bromopyruvate-treated enzyme from fraction 24 of the 5.0 S peak was not inhibited by phenylalanine under identical conditions. Similar experiments were performed with chorismate mutaseprephenate dehydratase from strain SA 34. It has been shown that both activities of the bifunctional enzyme from this strain Chorismate Mutase-Prephenate Dehydratase Vol. 246,No. 19 are completely insensitive to inhibition by 4 mar phenylalanine (3). The enzyme activity profiles in Fig. 6  The remaining chorismate mutase activity (50% of the starting activity) was completely desensitized to inhibition by phenylalanine.
A O.l-ml aliquot of this preparation was mixed with 0.09 mg of yeast alcohol dehydrogenase and was applied to a 5 to 20% sucrose gradient containing 0.10 M potassium phosphate (pH 7.4) and 5 PM phenylalanine. The gradient contained 0.34 unit of chorismate mutase activity. With this amount of enzyme the untreated control preparation was fully dimerized under these conditions. Centrifugation was for 10 hours at 46,000 rpm in a Spinco SW 50 rotor at 4". To 5 to 20% sucrose gradient.s containing 0.05 M potassium phosphate (pH 7.4), was added 1.0 to 1.6 units of native chorismate mutase-prephenate dehydratase or 5.4 units of feedback-insensitive enzyme from strain SA 34. Phenylalanine when present in the gradients was 1 mM. Yeast alcohol dehydrogenase (70 pg) was added to each enzyme sample and its position is indicated by an arrow (7.6 S). Other arrows designate the peak SPO.~ values for native enzyme with (O-O) and without (C-0) phenylalanine, and for feedback-insensitive enzyme with (a--+) and without (c+---•) phenylalanine. The figure is a composite of three separate gradients.
Centrifugation was at 37,000 rpm in an SW 39 rotor for 14 hours at 2' wit,h a Spinco L2-65 B centrifuge. SA 34 sedimented at 5.0 S whether phenylalanine was included in the gradient or not. The wild type enzyme in the absence of phenylalanine gave an intermediate sedimentation coefficient. This sedimentation coefficient was partially dependent on the extent of prior dialysis of the enzyme preparation to remove residual phenylalanine.
Even after extensive dialysis, however, the sedimentation coefficient of the wild type enzyme in these experiments could not be decreased below 5.8 S (Fig. 6). The molecular weight of the feedback-insensitive enzyme in the presence of 1.0 mu phenylalanine was estimated by gel filtration to be 120,000 (data not shown) which is the same as that obtained for the wild type enzyme without phenylalanine (Fig. 2). Neither tyrosine nor tryptophan had any effect on the sedimentation coefficient of the feedback-insensitive enzyme. Thus, the feedback insensitive enzyme does not dimerize in the presence of phenylalanine.
At pH values below 6, chorismate mutase activity was relatively insensitive to inhibition by phenylalanine whereas prephenate dehydratase was fully inhibited (2). The ability of phenylalanine to cause dimerization of the enzyme was studied at different pH values. The data in Fig. 7 show that at pH 6.0 10 pM phenylalanine was required to promote maximal dimer formation while at pH 7.4,5 PM phenylalanine was sufficient.
At pH 5.0 the comparison was even more striking.
At pH 5.0 maximal dimer formation required 100 times more phenylalanine than was needed at pH 7.4. In addition, even with 1 InM phenylalanine the estimated sedimentation coefficient was lower at pH 5.0 compared to that obtained at pH 7.4. These data show clearly that at low pH chorismate mutase-prephenate exhibits a decreased ability to dimerize in the presence of phenylalanine.
The experiments of Figs. 5 to 7, taken together, show that treatments that desensitize chorismate mutase-prephenate dehydratase to feedback inhibition by phenylalanine also reduce or prevent enzyme dimerization in the presence of inhibitor. The phenylalanine regulatory sites are thus implicated in dimerization.

Kinetics and Feedback Inhibition
of Prephenate Dehydratase Activity at pH 6.OSince decreased phenylalanine inhibition of chorismate mutase activity at pH 6.0 (2) was accompanied by The data at pH 7.4 are obtained from Fig. 3 and only points representing the mean of multiple det,erminations are shown. Other points are from single experiments except that at pH 5.0 the data point for 1000 PM phenylalanine is the mean of duplicate determinations.
Centrifugation was as described for Fig. 3. J. C. Xchmit and H. Zalkin decreased tendency of the enzyme to associate (Fig. 7)) we investigated the effect of low pH on the substrate saturation kinetics and inhibition of prephenate dehydratase by phenylalanine. Previous results (2) had indicated that prephenate dehydratase activity retained sensitivity to inhibition by phenylalanine at pH 6.0. As shown by the data in Fig. 8, prephenate dehydratase activity exhibited increased sensitivity to inhibition by phenylalanine at pH 6 relative to pH 7.4. At a phenylalanine concentration of 0.05 mM there was a 3-fold difference in sensitivity to feedback inhibition of prephenate dehydratase.
Replots of the data in Fig. 8 according to the modified Hill equation (47) indicated a decreased interaction coefficient, n', from 1.8 at pH 7.4 to 1.5 at pH 6.0. This effect would be expected to be more pronounced at an even lower pH (Fig. 7) but such experiments could not be performed because of the instability of prephenate below pH 6.0. We therefore conclude that at pH 6, under conditions less favorable to enzyme dimerization, phenylalanine binding is retained but interaction between sites is somewhat decreased.
The effect of low pH on prephenate dehydratase activity is shown in Fig. 9. Lowering the pH decreased the u,,, from 4 enzyme units at pH 7.4 to 3 enzyme units at pH 6.0 but did not appreciably alter the K, for prephenate.
In both cases normal kinetics was obtained in the absence of phenylalanine.
In the presence of 80 PM phenylalanine, cooperativity for prephenate is indicated by concave upward curvature in the double reciprocal plots. However, the cooperativity at pH 6.0 is less than that at pH 7.4. The interaction coefficient (n) for prephenate decreased from 2.3 at pH 7.4 to 1.6 at 6.0. Thus, the following changes occurred at pH 6.0. (a) Chorismate mutase activity was desensitized to inhibition by phenylalanine.

Separation of Monomeric and Dimeric Forms of Chorismate Mutase-Prephenate
Dehydratase-Under special conditions multiple peaks of chorismate mutase-prephenate dehydratase activity corresponding to monomeric and dimeric species were obtained by sucrose gradient centrifugation.
In order to obtain separation of monomeric and dimeric forms of the enzyme, low protein concentrations similar to those used for the experiments shown in Fig. 4 were required together with saturating concentrations of phenylalanine (100 pM or higher). It was also necessary to dialyze the enzyme prior to centrifugation in order to remove phenylalanine that was used to stabilize the enzyme during storage. After dialysis phenylalanine was added back to give a final concentration of 0.1 to 1.0 mu. The data in Fig. 10A show the activity profiles for chorismate mutase and prephenate dehydratase activities.
Monomeric and dimeric forms contained both activities.
The relative proportion between the two enzyme forms was influenced by the concentration of enzyme applied to the gradients. This is shown by a comparison of Fig. 10, A and B. For the experiment in Fig. lOA, 0.016 unit of chorismate mutase activity was applied in a volume of 0.05 ml to a gradient containing 100 PM phenylalanine.
From the chorismate mutase activity profile it was calculated that approximately 60% of the enzyme was dimeric and 40% was monomeric.
The same result was obtained with 1 mu phenylalanine.
For the experiment in Fig. 10B the same number of enzyme units was applied in a volume of 0.2 ml to a similar gradient.
Under these conditions the distribution was 25% dimer and 75% monomer. Fro. 9. Double reciprocal plots for substrate saturation of prephenate dehydratase at pH 7.4 and pH 6.0 in the presence and absence of phenylalanine.
Assays were similar to those described in Fig. 8 except that 0.025 unit of prephenate dehydratase was used and the prephenate concentration was varied in the presence or absence of 80 NM phenylalanine. for storage is shown by the experiment in Fig. 1OC. In this experiment the stock enzyme solution containing 1 mM phenylalanine was diluted to 0.32 unit per ml and 100 pM phenylalanine. An aliquot (0.05 ml) containing 0.016 unit of chorismate mutase activity was applied to a sucrose gradient containing 100 ELM phenylalanine.
Under these conditions following centrifugation 92% of the activity remained dimeric. The importance of prior dialysis to remove phenylalanine used Two distinct peaks of activity indicate that dimers and mono- mers sediment as discrete species. For this to occur, dimer once formed must be stable with little tendency to dissociate to monomer as long as phenylalanine is saturating. This expectation is supported by the results shown in Fig. 1OC. In this case the phenylalanine concentration was sufficient to maintain the enzyme in the dimeric form during dilution and centrifugation. Such a result requires that dissociation of the dimer to monomer must be very slow when the enzyme is saturated with phenylalanine.

Dimerixation
Is Not Obligatory for Inhibition-Dimerization has been shown to be dependent on the inhibitor, phenylalanine, and appears to involve the allosteric inhibitor binding site(s). However, dimerization is not obligatory for inhibition of either activity.
Under conditions in which two peaks of enzyme activity (monomer and dimer) were obtained on sucrose gradients, the monomeric chorismate mutase-prephenate dehydratase activities were strongly inhibited by phenylalanine (Fig. 10A). Inhibition of chorismate mutase activity by phenylalanine is not shown in Fig. 10A for reasons of clarity.
The enzyme would not be expected to dimerize during the assa,y because it was diluted considerably during centrifugation and was diluted still further (1: 5) in the assay mixture.
Such dilution would favor dissociation rather than association.
In control experiments, both enzyme activities were strongly inhibited by phenylalanine at 22" and at 2-4". Furthermore, neither substrate promoted dimer formation.
The enzyme should therefore remain monomeric during assay. It is thus concluded that the monomeric species is subject to inhibition by phenylalanine.

DISCUSSION
The data reported in this study confirm previous results (2) which show that the sedimentation coefficient of chorismate mutase-prephenate dehydratase is increased in the presence of phenylalanine.
Sucrose gradient centrifugation in conjunction with gel filtration was used to distinguish between a large conformational change and an association-dissociation of the enzyme. The large increase in sedimentation coefficient of chorismate mutase-prephenate dehydratase caused by phenylalanine could be due either to association (i.e. increase in molecular weight) or to a large conformational change resulting in a smaller more compact molecule.
Conformational cha,nges of this magnitude have been reported (48). Gel filtration which separates proteins based on Stokes radius or size (37, 38) can be used to distinguish between these two possibilities.
If the increase in sedimentation coefficient caused by phenylalanine is due to enzyme association, gel filtration should reveal an increased molecular weight.
If the increase in sedimentation coefficient caused by phenylalanine is due to a conformational change, giving a more compact and therefore smaller molecule, the enzyme should be retarded during gel filtration and give a decreased estimated molecular weight.
It is clear from the data in Fig. 2 that phenylalanine increased the apparent molecular weight of the enzyme as determined by Sephadex G-200 gel filtration. Therefore, phenylalanine caused an association of the enzyme. This is further supported by the observation that in the presence of phenylalanine the sedimentation coefficient was dependent on enzyme concentration (Fig. 4). Theobservedmolecular weight change is consistent with a rapid monomer-dimer transition.
Gel filtration experiments show that the estimated molecular weight is increased 2-fold in the presence of phenylalanine (Table I). Molecular weight estimations of native enzyme by sucrose gradient centrifugation indicate a 1.7fold increase in the presence of phenylalanine.
Although molecular weight estimation by sucrose gradient centrifugation can provide only an approximation of true molecular weight, this deviation from a doubling may be significa,nt.
It was consistently observed that the molecular weight of feedback-insensitive enzyme estimated by sucrose gradient centrifugation was lower than that of the native enzyme determined in the absence of phenylalanine (Fig. 6). One possibility is that the feedback-insensitive enzyme was completely monomeric because of inability to associate whereas a fraction of the native enzyme was in rapid associationdissociation perhaps due to tightly bound phenylalanine or a tendency to dimerize in the absence of inhibitor.
The reason for the large difference between molecular weights estimated from gel filtration and sucrose gradient centrifugation experiments is not fully understood.
A contributing factor is that yeast alcohol dehydrogenase which was used as a reference for sucrose gradient centrifugation did not fall on the straight line obtained from the other four standard proteins in gel filtration experiments (Fig. 2). However, this complication does not detract from the important conclusion that phenylalanine caused chorismate mutase-prephenate dehydratase to dimerize. A higher concentration of phenylalanine was required to inhibit either chorismate mutase or prephenate dehydratase activity (2) than was needed for maximal association at relatively high enzyme concentration (Fig. 3). For 50% inhibition of chorismate 'mutase or prephenate dehydratase activities (assayed with 0.2 mm chorismate or 0.9 mM prephenate, respectively) 100 to 200 PM phenylalaninc was required whcrcas maximal enzyme association, determined in the absence of substrates, was obtained with approximately 5 PM phenylalanine. It is possible that this discrepancy is due, at least in part, to the competitive relationship between phenylalanine and substrates. A close correlation between dimerization and feedback inhibition is indicated by the observations that compounds which inhibited both activities caused the enzyme to dimerize (Table I) and that desensitized enzyme did not dimeriae (Figs. 5 and 6). Tyrosine, which does not inhibit either activity, may be a possible exception (Fig. 3). The concentration of tyrosine required to cause dimerization was 100 to 500 times higher than that required for a comparable effect with phenylalanine.
It is possible that tyrosine is a weak analogue of phenylalanine and its effect on the enzyme, although not strong enough to cause inhibition in the presence of substrates, is sufficient to cause some dimerization in sucrose gradients.
Enzyme desensitized to phenylalanine inhibition either by bromopyruvate treatment or by genetic modification did not dimerize in the presence of feedback inhibitor (Figs. 5 and 6). This suggests that binding of phenylalanine to the regulatory site(s) is responsible for both inhibition and dimerization.
The effect of low pH on association is complex. At low pH, relative to pH 7.4 considerably more phenylalanine is required to promote dimerization.
This cannot result from a weaker affinity of phenylalanine for the enzyme because the sensitivity of prephenate dehydratase to feedback inhibition is maintained (Fig.  8). It is also unlikely that there are two classes of phenylalanine sites that participate in association or inhibition (or both) because association cannot occur in the feedback-insensitive enzyme from mutant strain SA 34. It therefore appears possible that at low pH a phenylalanine-induced change that leads to inhibition of prephenate dehydratase camlot be transmitted to the chorismate mutase site. These experiments show two main points : (a) concomitant alteration of feedback inhibition of one enzyme activity and phenylalanine-induced dimerization, and (b) decreased homotropic and heterotropic cooperative interactions under conditions of decreased enzyme association.
Although the results of several experiments indicate a relationship between feedback inhibition and dimerization, this effect is not obligatory for inhibition of either catalytic activity by phenylalanine.
At sufficiently low protein concentration a major fraction of chcrismate mutase-prephenate dehydratase remained monomeric and yet both activities were sensitive to inhibition by phenylalanine (Fig. 10A). Any model proposed to explain assoication-dissociation of enzyme and the relationship between different forms with active and inactive states must accommodate four experimental observations.
(a) A single apparently symmetrical peak of activity with intermediate sedimentation coefficient is obtained with rela-2A FIG. 11. Hypothetical scheme for association-dissociation and inhibition of enzyme activity by phenylalanine (phe). A and AA represent monomeric and dimeric forms of active chorismate mutaae-prephenate dehydratase.
Z and ZZ represent monomeric and dimeric forms of the inhibited enzyme. All species except AA have been detected by sucrose gradient centrifugation.
The dashed U.TTOZL indicates a questionable step. Reaction of ZZ to II does not occur to a significant extent. tively high enzyme concentration and less than saturating concentrations of phenylalanine.
Therefore, under such conditions a rapid association-dissociation occurs. (6) A similar rapid association-dissociation occurs under conditions of relatively low enzyme concentration and less than saturating concentrations of phenylalanine.
(c) Distinct monomer and dimer species are observed at relatively low enzyme and saturating phenylalanine concentrations.
Therefore, under the latter conditions rapid association-dissociation does not occur. (d) Association-dissociation is phenylalanine-and protein concentration-dependent. A simple hypothetical scheme that accommodates these observations is shown in Fig. 11. All species except active dimer (AA) have been detected.
It is proposed that active monomers (A) are converted into inhibited monomers (I) in the presence of phenylalanine in a reversible step. Inhibited monomers may dimerize (II) in a concentration-dependent step. Dissociation of inhibited dimer to inhibited monomer may not occur readily. Upon dissociation of phenylalanine an active dimer may be formed which dissociates into active monomers.
Such a cyclic scheme appears necessary in order to account for a rapid association-dissociation equilibrium at low phenylalanine concentration and a very slow association at low protein but saturating phenylalanine concentration.
It is proposed that monomeric and dimerit species are observed under the latter conditions because low enzyme concentration ensures that association will be incomplete and high phenylalanine concentration prevents dissociation of inhibited dimer to active monomers via active dimer.
In this report the smallest active species of molecular weight approximately 88,000 to 109,000 has been designated a monomer. Preliminary experiments by gel electrophoresis in the presence of sodium dodecyl sulfate and mercaptoethanol (49) suggest that the monomer is composed of two polypeptide chains of similar size. The enzyme is the product of the phe A locus in X. typhimurium (50).