Evidence against a covalent intermediate in the adenosine triphosphate phosphoribosyltransferase reaction of histidine biosynthesis.

14C-Labeled 5-phospho-alpha-D-ribose-1-diphosphate (PRibPP) was synthesized and its interaction with adenosine triphosphate phosphoribosyltransferase was examined by gel filtration in a search for a form of this substrate covalently bound to the enzyme. Wide variation in solvent conditions gave little labeling of the enzyme. Heavy labeling was found only in the presence of the second substrate, ATP, and this was shown to arise from tightly but noncovalently bound product. Previous reports of a covalent intermediate in this enzymatic reaction probably were due to contaminating ATP in 5-phospho-alpha-D-ribose-1-diphosphate. Feedback inhibition of the enzyme by histidine was shown to occur at the step giving product or at some earlier step in the mechanism.


WAYNE T. BRASHEAR AND STANLEY M. PARSONS
From the Department of Chemistry, University of California, Santa Barbara, California 93106 Y-Labeled 5-phospho-a-n-ribose-1-diphosphate (PRibPP) was synthesized and its interaction with adenosine triphosphate phosphoribosyltransferase was examined by gel filtration in a search for a form of this substrate covalently bound to the enzyme. Wide variation in solvent conditions gave little labeling of the enzyme. Heavy labeling was found only in the presence of the second substrate, ATP, and this was shown to arise from tightly but noncovalently bound product. Previous reports of a covalent intermediate in this enzymatic reaction probably were due to contaminating ATP in 5-phospho-cu-n-ribose-l-diphosphate. Feedback inhibition of the enzyme by histidine was shown to occur at the step giving product or at some earlier step in the mechanism.
Histidine is synthesized in microorganisms via a pathway composed of 10 enzymatic steps which is subject to two forms of regulation (1). Feedback inhibition of the first enzyme provides a rapid response to fluctuations in the histidine pool (2, 3) while repression control of the enzyme levels provides long term control.
The first enzyme of the pathway is adenosine triphosphate phosphoribosyltransferase (EC 2.4.2.17), abbreviated ATP phosphoribosyltransferase, 1 and it catalyzes the reaction show in Equation 1.' ATP phosphoribosyltransferase is a hexamer composed of 36,000 molecular weight subunits (4) which can bind histidine at an allosteric site to give inhibition of the enzymatic reaction (3). In early work on the enzyme, Martin (2) observed the exchange reactions shown in Equations 2 and 3.
The existence of exchange reactions when only one of the two substrates was present, for the reaction in either direction, suggested that an enzyme covalent intermediate was formed.
* This work was supported by Grant NIH-LROl-GM 20835 from the United States Public Health Service. 'Adenosine triphosphate phosphoribosyltransferase has been incorrectly termed PR-ATP synthetase in some of the literature. The full systematic name is N'-(5'-phospho-@-n-ribosyl)ATP:pyrophosphate phosphoribosyltransferase.
The reaction product solution always was immediately gel-filtered on a column (0.9 x 20 cm) of Sephadex G-50 at 25" in a buffer identical MATERIALS AND METHODS ATP phosphoribosyltransferase was isolated as previously described (12) from Salmonella typhimurium LT2 strain TA2165 (13) or by a newer procedure (13) from strain hi&l1 (14).  The ATP phosphoribosyltransferase substrate analogs AMP and ADP were tried. Both bind to the enzyme but are not converted to products (2,19). No labeling was found after reaction in the presence of AMP and little was found after the presence of ADP (Table I,  Experiments 24 and 25). Thus "substrate synergism" could not be demonstrated using AMP and ADP.
was preincubated for 50 min in the presence of histidine (Fig.  2), but only slightly blocked if it was preincubated for only 7 s in the presence of histidine (Fig. 3). This latter behavior is similar to the lag in histidine inhibition seen for enzyme which has not been exposed recently to histidine (3,18). The high level of labeling in the presence of ATP thus responds to histidine inhibition in the same manner as steady state formation of products. Also, a trace level of ATP added 7 s after addition of histidine gave little labeling (Table I, Experiment 26). Thus, stoichiometric and not catalytic levels of ATP are required for extensive labeling. Finally, it was found that histidine was not required at all for labeling since reaction under the conditions of Fig. 1 in the complete absence of histidine led to 75% labeled enzyme (Table I, Experiment  27). However, when excess ATP was included, a very high level of Enzyme which had been highly labeled was isolated by gel apparent labeling was found (Fig. 1). Labeling did not depend filtration as in Fig. 1, denatured with urea, and gel-filtered in upon order of addition of ATP and ['"C]PRibPP (not shown). urea to determine whether radioactivity was still bound to the This extensive labeling was completely blocked if the enzyme protein. Fig. 4 shows that it was not. Also, the labeled native  2 (left center). Effect of prolonged incubation with histidine on labeling.
The enzyme was treated as in Fig. 1  conditions similar to the first experiment, except that no histidine was added before urea denaturation, no radioactivity was bound to the enzyme after gel filtration. This is in contrast to a previous report (5). Thus, no evidence was obtained that the high level of labeling in the presence of both substrates was of a covalent nature.
These results made it probable that the radioactive material bound to the enzyme was the product ['"C]PRibATP. This was tested by incubating the enzyme in the presence of labeled PRibPP, histidine, and ATP under the conditions of Fig. 1, with the subsequent addition of a partially purified preparation of the second and third enzymes of the histidine pathway genetically designated E and I. These enzymes convert PRib-ATP to PRibAMP and BBM II (1). Gel filtration of the reaction product showed 90% diminished labeling of ATP phosphoribosyltransferase compared to the control (not shown). Also, a heavily labeled enzyme solution similar to that gel-filtered in Fig. 1 was analyzed by paper electrophoresis. A predominant new peak of radioactivity having a similar electrophoretic mobility to authentic PRibATP was present along with excess ['*C]PRibPP and a diffuse peak near the origin probably arising from [14C]PRibATP bound to the enzyme (Fig. 5). Thus the radioactive label behaved like PRibATP.
Since extensive labeling of ATP phosphoribosyltransferase by [14C]PRibPP could not be demonstrated under any set of conditions tried unless ATP was added, it seemed probable that the low level of labeling observed with [**C]PRibPP was due to contaminating ATP. To test this, an enzyme solution reacted with ['"C]PRibPP was treated with E and I enzymes. Gel filtration of the reaction product showed no labeling of ATP phosphoribosyltransferase (Fig. 6B), whereas a control without E and I enzymes showed 0.2% labeling (Fig. 6A). When the labeled PRibPP was pretreated with hexokinase and glucose before incubation with ATP phosphoribosyltransfer- in the absence of ATP. These experiments were conducted with enzyme in standard buffer without glycerol at pH 7.5. Enzyme (500 ~1, about 32-nmol subunits) was incubated 60 min at 25' with 0.5 unit of pyrophosphatase, followed by the addition of ["C]PRibPP (14 ~1, 40 nmol containing 43,000 cpm) and incubation for 7 s, followed by the addition of histidine (4 ~1, 400 nmol) to give 0.8 mM histidine. To 250 ~1 of this solution 25 ~1 of partially pure E ase, labeling of ATP phosphoribosyltransferase again was much reduced (Fig. 6C) (5). The experimental conditions explored included those in which the enzyme exists in two forms of the hexamer and a high molecular weight aggregated form (11). Enzyme isolated using two different procedures (12,13) gave the same results. A high level of labeling could not be reproduced under any environmental variation. Heavy labeling could be obtained only in the presence of high concentrations of ATP. This labeling was shown to be capable of transformation to ["ClPRibATP since it largely was eliminated by other enzymes of the histidine pathway which act on PRibATP, and most of the label possessed the electrophoretic mobility of PRibATP.
In addition, no label remained bound to the enzyme after urea denaturation, thus demonstrating that the label bound to native enzyme was not linked by a stable covalent bond. There was no evidence that the and I enzymes was added. The solution was incubated 30 min at 25", and gel-filtered in the standard buffer plus 1 mM histidine. This gave Graph B. The remaining solution, untreated by E and I enzymes, then was immediately gel-filtered and gave Graph A.
["C]PRibPP treated with hexokinase and glucose as detailed under "Materials and Methods" was used for labeling 200 ~1 of enzyme using the procedure of Graph A, except that the product was gelfiltered immediately after quenching with histidine. The result is shown in Graph C. Decreased labeling seen in B and C suggests that the labeling in A is due to PRibATP. reaction. Tight binding of PRibATP also could cause the apparent differences between pure enzyme and enzyme in crude bacterial extracts reported previously (13). Bound PRibATP from a crude extract probably would stabilize the hexameric form of the enzyme (11). This would make crude enzyme resistant towards cold inactivation and would change the precipitation behavior of the enzyme in Ouchterlony immunodiffusion tests as was observed.
Many at the step involving reaction between PRibPP and ATP, or at some earlier step. Histidine inhibition due to slowed product release from the enzyme clearly is excluded as a dominant mechanism for inhibition of the steady state reaction since not even enzyme levels of product are formed under these conditions. Also, histidine was demonstrated to be not an effective inhibitor of PRibATP formation within 7 s after being mixed with the enzyme. This result confirms previous observations that enzyme which has not been recently exposed to histidine requires several minutes before steady state activity is completely inhibited upon addition of histidine (3,19).