Stereochemical course of the adenosine triphosphate phosphoribosyltransferase reaction in histidine biosynthesis.

The product of the first reaction in histidine biosynthesis is shown by optical rotation measurements on three derivatives to have inverted, beta stereochemistry at the newly formed bond. This is in contrast to alpha linkage expected on the basis of previously observed exchange, specificity, and covalent intermediate phenomena. The postulate double displacement mechanism for adenosine triphosphate phosphoribosyltransferase must be modified to account for the product stereochemistry.


RESULTS
Before proceeding with optical rotation measurements and their interpretation it was important to verify the proposed structure of PRibATP, which is illustrated in Fig. 1. The proton magnetic resonance spectrum is given in Fig. 2. Distinctive features are as follows. The two downfield singlets at 8.37 and 8.25 ppm are the H-8 and H-2 protons of the charged adenine moiety, respectively (18). The 2 C-l protons of the ribose rings can be recognized as doublets centered at 5.78 and 5.60 ppm. The two C-5' methylene groups appear at 3.76 and 3.51 ppm. The remaining ribose ring protons resonate in an envelope between 4.0 and 3.6 ppm. Also the respective chemical shifts are similar to those of l-methyladenosine as shown in Fig. 3. Thus the proton magnetic resonance spectrum of PRibATP confirms its proposed structure. The ultraviolet spectrum of PRibATP exhibits a pH-dependent titration centered at pH 8.8 (13). At higher pH the spectrum exhibits X,,, at 260 nm, a shoulder at 265 nm, and a plateau from 285 to 300 nm. At lower pH it more closely resembles protonated ATP, with a X,,, at 260 nm. No adequate theoretical approach is available yet for predicting the optical rotary behavior of this chromophore depending on the configuration of two intimately bonded asymmetric centers (19). The new absorption bands, whose transition moments are not characterized, would give unpredictable Cotton effects arising from the asymmetric perturbation of the chromophore by the two ribosyl moieties. Since these Cotton effects occur at high wavelength they would dominate rotational behavior in the near-ultraviolet and visible. In fact, the molecular rotation of PRibATP measured at the sodium D-line doublet changes from positive to negative depending upon protonation of the chromophore, clearly demonstrating complication from nearultraviolet Cotton effects (Table II) The strategy adopted here to determine the unknown stereochemistry was to move the asymmetric center in question which did not affect its stereochemistry. The van't Hoff principle of optical superposition (20) could then be applied. This principle assumes that molecular rotation of a molecule measured at the sodium D-line will be equal to the sum of rotations separately contributed by each asymmetric unit. The principle can be valid if the asymmetric units do not strongly interact with each other (19). This was accomplished in two quite different ways, and with a third related measurement as as far away as possible from the chromophore in a manner follows.  Ames et al. (13), and is illustrated in Fig. 4. Since the intermediate phosphoribosylaminoglycoside transiently formed during this rearrangement is conjugated to the aglycone it is not expected to be subjected to facile mutarotation at the C-l" center as is the case with an unconjugated amine (23). Thus the stereochemistry at C-l" will be unchanged in rearranged PRibATP.
In addition to moving the C-l" assymmetric center further away from the adenine moiety the rearrangement allows increased torsional motion of C-l" with respect to the adenine ring. Both changes in the relationship of C-l" to the adenine ring will minimize net interactions between them (19). The asymmetric units of rearranged NB-PRibATP can be thought of as being similar to ATP and N-glycylribosylamine (Table I). Both anomers of N-glycylribosylamine have been synthesized by others and characterized, as well as the carbobenzoxy derivatives. Phosphorylation of the 5' position of ribosides has little effect on rotation (19). Summing the molecular rotation for either the a-or @-ribosylamine with that for ATP (26) yields the two predictions for (Y or p C-l" linkage in rearranged N6-PRibATP given in Table I was partially complete rotation in the visible changed from positive to negative depending upon wavelength. Final molecular rotation at the sodium D-line was -260", in good agreement with the prediction for the 0 configuration of C-l" (Table II).
To eliminate the possibility of any special conformational  (23) + 127" (23) a References are given in parentheses. -258" -96" (24) -161' (25) by guest on March 24, 2020 http://www.jbc.org/ Downloaded from  interactions which would invalidate the optical superposition principle, rearranged NG-PRibATP was oxidized with periodate. Periodate oxidized cy or P-N-glycylribosylamine and adenosine serve as models for this product ( Table I). The molecular rotation of oxidized rearranged NG-PRibATP of -330" also is in good agreement with the prediction for p stereochemistry at C-l" (Table II). An independent method for confirming the stereochemistry is available by utilizing the metabolic pathway itself. PRi-bATP can be converted to BBM II by incubation with a bacterial extract which contains E and I enzymes but which is blocked mutationally at the A step (Fig. I). Both E and I steps are hydrolyses requiring no cofactors other than magnesium. BBM II can be thought of as being composed of either cy or /3-N-glycylribosylamine and aminoimidazolecarboxamide ribonucleotide, a well characterized intermediate of purine biosynthesis (Table I). Incubation of PRibATP with E and I enzymes gave a product with molecular rotation of -400" in fair agreement with the prediction for /l stereochemistry at C-l" (Table II). Thus all three measurements indicate p stereochemistry in the newly formed bond of PRibATP.

DISCUSSION
The generally good agreement between observed and predicted rotations for all p linkage in PRibATP is strong support for inversion of configuration at the newly formed linkage. Since o( linkage should have yielded positive rotations, it is highly unlikely that any special conformational effects accidentally would cause inversion of sign to give negative rotation for all three measurements. The occurrence of the p configuration in PRibATP makes it very unlikely that it arose from cu-PRibPP by a double displacement mechanism. Chemical studies on ribofuranosides give no evidence for front-side displacement reactions at C-l (27). Thus the simple double displacement mechanism for ATP phosphoribosyltransferase must be modified.
There are at least three reaction mechanisms, shown in Fig.  6, that can account for the exchange reactions, formation of a phosphoribosylated enzyme, and the p stereochemistry of product. These are the simplest mechanisms which incorporate essential elements characteristic of other enzymatic glycosyl reactions, such as those of lysozyme (28), sucrose phosphoryl ase (29), @-galactosidase (30), other phosphoribosyltransferases (31), and purine nucleoside phosphorylase (32). Possible mechanisms which invoke anchimeric assistance from the substrate itself are unlikely since the C-2 hydroxyl of PRibPP is cis to the leaving group and the C-5 phosphate can attack C-l only in a strained conformation.
A triple S,2 displacement could occur involving two phosphoribosylated enzyme intermediates both on the normal pathway. Pyrophosphate might exchange with the first intermediate and ATP with the second. Histidine could bind to either one or both intermediates to stabilize them (15). Alternatively, the apparent phosphoribosylated enzyme intermediate might not be on the pathway, but could be trapped in a reversible side reaction by histidine. The normal enzymatic pathway then could involve either a single S N2 displacement, as shown on the second line of Fig. 6, or a carbonium ion intermediate, as shown on the third line, which could give either N-or P-PRibATP.
In the latter two cases reversible side reactions also would have to occur in the presence of PRibATP in order to account for exchange reactions.
Steady state initial velocity rate determinations of enzymatic activity as a function of substrate and histidine could be useful in limiting mechanistic possibilities. However, previous studies are very contradictory (2, 6, 7, 10) as a result of uncharacterized conformational and dissociative equilibria of ATP phosphoribosyltransferase (5) and contaminating histidase (4). Rate studies have not been performed on the exchange reactions. More information will be required to decide among the possibilities in Fig. 6, in addition to the possibility that some of the previous data were incomplete or incorrect. Observation of inversion in the adenosine triphosphate phosphoribosyltransferase reaction as shown here indicates that exchange, specificity and "covalent intermediate" criteria (15)