Purification and Properties of the Wild Type and a Feedback-resistant Phosphoribosyladenosine Triphosphate

A purification procedure has been devised for phosphoribosyltransferase, the L-histidine-sensitive first enzyme in the pathway for histidine biosynthesis in Salmonella fyphimurium. The procedure was applied to a wild type and a feedbackresistant strain. The enzymes from both strains appeared nearly homogeneous in the ultracentrifuge and upon polyacrylamide gel electrophoresis in urea and sodium dodecyl sulfate. The enzymes had similar sedimentation coefficients in the ultracentrifuge and similar mobilities on polyacrylamide gels containing sodium dodecyl sulfate. Tryptic peptide maps of the two enzymes could not be distinguished. The wild type enzyme gave regular Michaelis-Menten kinetics but initial velocity analysis at a constant optimal magnesium to ATP ratio (2: 1) gave nonparallel lines on double reciprocal plots. L-Histidine was an uncompetitive inhibitor with respect to phosphoribosyl pyrophosphate, while it was a noncompetitive inhibitor with respect to ATP. The curves for L-histidine and L-thiazolealanine inhibition were sigmoid in shape, and conversion to Hill plots gave straight lines with slopes of 1.6 and 1.8, respectively. Inhibition by both effecters was pa-dependent. The reverse reaction was also inhibited by L-histidine. A difference spectrum of the wild type enzyme showed a striking increase in absorbance at 280 rnp upon the addition of L-histidine, whereas that of the feedback-resistant enzyme remained constant following the addition of L-histidine.

From the Section on Microbial Genetics, National Institute of Arthritis and Metabolic Diseases, National Institutes of Health, Bethesda, &larylancl 20014 SUMMARY A purification procedure has been devised for phosphoribosyltransferase, the L-histidine-sensitive first enzyme in the pathway for histidine biosynthesis in Salmonella fyphimurium. The procedure was applied to a wild type and a feedbackresistant strain.
The enzymes from both strains appeared nearly homogeneous in the ultracentrifuge and upon polyacrylamide gel electrophoresis in urea and sodium dodecyl sulfate.
The enzymes had similar sedimentation coefficients in the ultracentrifuge and similar mobilities on polyacrylamide gels containing sodium dodecyl sulfate.
Tryptic peptide maps of the two enzymes could not be distinguished.
The wild type enzyme gave regular Michaelis-Menten kinetics but initial velocity analysis at a constant optimal magnesium to ATP ratio (2: 1) gave nonparallel lines on double reciprocal plots. L-Histidine was an uncompetitive inhibitor with respect to phosphoribosyl pyrophosphate, while it was a noncompetitive inhibitor with respect to ATP. The curves for L-histidine and L-thiazolealanine inhibition were sigmoid in shape, and conversion to Hill plots gave straight lines with slopes of 1.6 and 1.8, respectively. Inhibition by both effecters was pa-dependent.
The reverse reaction was also inhibited by L-histidine. A difference spectrum of the wild type enzyme showed a striking increase in absorbance at 280 rnp upon the addition of L-histidine, whereas that of the feedback-resistant enzyme remained constant following the addition of L-histidine.
N-1-(5'-Phosphoribosyl) adenosine triphosphate:pyrophosphate phosphoribosyltransferase (EC 2.4.2) catalyzes the initial committed reaction in the L-histidine biosynthetic pathway of Salmonella typhinzurium (1). Mg++ The gene (hisG) that codes for phosphoribosyltransferase activity is located in a cluster of nine genes on the S. typhimurium chromosome. These genes specify the structures of all of the enzymes for histidine biosynthesis.
The genes are clustered in the order operator G-D-C-B-H-A-F-I-E. Regulated as a unit, they constitute the histidine operon (see review by Ames et al. (2)). The phosphoribosyltransferase is of interest because it is subject to feedback inhibition by the end product of the pathway, L-histidine (1). Martin (3) has conducted a detailed study of the wild type phosphoribosyltransferase with special emphasis on the feedback inhibition mechanism, with partially purified preparations of the enzyme. The wild type enzyme has recently been purified (4) to near homogeneity. Sheppard (5) has isolated a series of mutants that are resistant to the histidine analogue, thiazolealanine, and have lesions lying in the G gene. Each of these thiazolealanine-resistant mutants contains a feedback-resistant phosphoribosyltransferase.
In order to elucidate any physical or chemical differences between the wild type and feedback-resistant enzymes, the two proteins have been purified to near homogeneity with a new purification scheme. This purification procedure, the kinetic and feedback properties of the purified wild type enzyme, and a comparison between the two proteins form the subject of this report. MATERIALS AND METHODS

Bacterial Strains
The wild type phosphoribosyltransferase was isolated from strain hisTl504 hisC146,* which contains a histidine-constitutive mutation, hisT1504 (7), and a frameshift mutation, hisC146 (8). This strain was constructed by transducing hisCl46 into strain hisTl504 hisOGDCBH2253 (9). The phage mutant L4 of P22 (10) was used in the transduction, and auxotrophic recombinants capable of growth on histidinol (D gene function) were selected. A nonlysogenic recombinant was picked and used as a source for the wild type enzyme.3 The feedback-resistant phosphoribosyltransferase was isolated from strain hisGIl hisIF (5).

Growth of Bacterial Strains
('ells were cultured in a 300.liter fermenter at 37" and harvested in the late log phase of growth by centrifugation. 4 Nutant hisTl504 hisCf46 was grown in Medium E of Vogel and Bonner (11) containing 0.1 mu L-histidine.
Use of the sodium citrates in Medium E as the sole carbon source resulted in a 30.fold derepression of the phosphoribosyltransferase in constitutive cells as compared to wild type strain LT2 cells grown under the same conditions. Cells were grown in the fermenter from a starting inoculum of 500 cc. With forced aeration, mutant hisTl504 hisCl46 grew with a generation time of approximately 80 min in the fermenter.
The yield of late log phase cells after 16 hours of growth was 400 to 500 g (wet weight) of cells.
The feedback-resistant strain, hisGIl hisIFlS5, was grown in the fermenter on Medium E (11) containing 0.5% glucose, 0.06 1x1~ histidinol, 0.0075 mM L-histidine, and 0.4 InM adenine. These conditions result in up to a 25-fold derepression of the histidine biosynthetic enzymes as compared to organisms grown on escess histidine (12). The fermenter yield of cells from a l-liter inoculum after 16 hours of growth was 900 g (wet weight) of cells.

Enzyme Assays
Two assays for the forward reaction were used in this study. CoCpled G-70 Assay-The coupled assay is based on the con-2 The nomenclature is that of Dr. P. E. Hartman. The letters designate the gene(s) affected by a mutation and are followed by a mutat,ion number. Mutations in the T gene, which lie outside of the histidine operon, are constitutive for the histidine biosynthelir enzymes (7). The 1100 series of mutants isolated by Hhepp;trd (5) contain feedback-resistant phosphoribosyltransferase enzyme.
3 Strain hid'146 contains a frameshift mutation which is a polar mutation (8). It was necessary to construct a strain carrying a polar mutat ion because Voll (personal communication) has shown that histidine-constitutive nonpolar strains produce an unidentified substance which retardsthe flow rates of column chromatography.
Kepeated attempts to construct an equivalent strain cont,aining a polar mutation and a histidine-constitutive mutation for the feedback-resistant, mutant failed. 4 The authors are indebted to I). Rogerson, I). Johnson, and .J. Hicks who carried out the fermenter runs. 5 T. KIopotowski, personal communication.
version of the product of the forward reaction, PR--kTP, to BBM III (1). BBM III has a,n extinction coefficient of 8.0 x lo3 M-l cm-l at 290 mp (12). dn extract of strain hisG70 provided an excess of the enzymes required to convert PR-ATP to BBYU III. This assay can be used for assaying any strain regardless of whether or not the strain has the enzymes for conversion of PR-ATP to BBM III (1). The reaction mixture, similar to that previously reported (4), consisted of 30 pmoles of Tria-HCl buffer (pH 8.5), 3 pmoles of i\IgC&, 45 pmoles of KCl, I .5 pmoles of ATP, 0.15 pmole of PPribose-P, 5 ~1 of an estract of strain hisG70 that had been passed through a Sephadex G-50 column (13), and enzyme in a final volume of 0.3 ml. To initiate the reaction, PP-ribose-P was added and the initial change in absorbance at 290 mp was followed on a Beckman spectrophotometer equipped with a Gilford multiple sample absorbance recorder.
When enzyme, either of the substrates, or magnesium chloride was omitted from the reaction mixture, no appreciable change in absorbance occurred.
One unit of activity is defined as an initial change in absorbance at 290 rnp of 0.10 per 5 min. Based upon the extinction coeficient of BBM III (12), 1 unit of activity corresponds to the formation of 0.75 nmole of BB?yl III per min. The assay is linear with time and enzyme concentration to an initial change in absorbance of 0.150 per 5 min at 290 mp.

Pyrophosphatase
Assay-The assay measures the formation of PR-ATP at 290 rnp and has been previously described (4). This assay was used in studies of the purified enzyme. The routine assay gave identical reaction rates regardless of whether or not the solutions of PP-ribose-P contained EDTA.
As previously defined (4), 1 unit of activity is an initial change in absorbance of 0.10 per 5 min and corresponds to the formation of 1.67 nmoles of PR-ATP per min. The units of enzyme activity obtained from this assay can be converted to units of enzyme activity fol the coupled G-70 assay by multiplying by 2.9. Assay of Reverse Reaction--The assay of the reverse reaction depends on the determination of the rate of decrease in absorbance at 290 rnp of PR-ATP.
Neither ATP nor PR-ATP absorbs appreciably at 290 1nl.L (1). The reaction mixture contained 10 pmoles of Tris-HU buffer (pH 8.5), 0.3 pmole of tetrasodium pyrophosphate, 6.8 nmoles of PR-ATP, 0.5 pmole of M&12, and enzyme in a total volume of 0.3 ml. PR-ATP was added to start the reaction, and the initial change in absorbance at 290 mp was followed.
The assay was linear with time and enzyme concentration to an initial chaqe in absorbance of 0.02 per 4 min. One unit of activity is defined as an initial change in absorbance of 0.1 per 4 min. This corresponds to the disappearance of 2.08 nmoles of PR-ATP per min. The enzymatic synthesis of PR-ATP is described below.

Polyacrylumiok Gel Electrophoresis
Electrophoresis of the enzyme in polyacrylamide gels containing urea was performed according to the method of Reisfeld and Small (14). Gels were stained for proteins with a 0.25% solution of Amido black (Allied Chemicals) in 7% acetic acid for 1 hour and were destained with 7% acetic acid.
The subunit molecular weight of the wild type and feedbackresistant enzymes was determined in 7.5% polyacrylamide gels containing SDS as described by Shapiro, Vifiuela, and Maize1 (15).

RESULTS
Difference absorption spectra were measured in a Cary model 15 recording spectrophotometer with cuvettes with a l-cm light path. Enzyme was prepared for these studies by overnight dialysis against 2 liters of the buffer described by Voll, Appella, and Martin (4) at pH 8.6 and containing 0.01 M P-mercaptoethanal. The temperature was 25".
Enzymatic Synthesis of PR-ATP and PR-AMP
Buffer A differs from the standard buffer by containing 0.1 M Tris-HCl, pH 7.5.

PR-ATP and PR-AMP
were generated in a cell-free extract of mutant hisTl504 hisWAFIE232'7 (9) by the method of Ames, Martin, and Garry (I) as modified by Smith and Ames (16). The PR-AMP and PR-ATP were separated from the crude reaction mixture by chromatography on DEAE-cellulose as described by Smith and Ames (16). The PR-ATP fraction was rechromatographed once. The purified PR-ATP had an A2e0: A290 ratio of 5.38 at pH 8.5, the PR-AOIP had an A2e0:AS9,, ratio of 10 at pH 8.5 (16).
Preparation of Extracts-All procedures were performed at 4" unless otherwise stated.
In a typical purification, 1000 g, wet weight, of cells were suspended in a minimal volume of Buffer A and homogenized in a Waring Blendor at half-minimal speed foi 5 min. Following homogenization, Buffer A was added to a total volume of 2 liters.
The cells were disrupted by two passages through a Branson sonifier, equipped with a 0.5-inch probe and a continuous flow attachment, at a current of 5 amp. The cell debris was removed by centrifugation at 27,000 x g for 70 min in a Servall RC-2B centrifuge.

LWacenfrijugation
Sedimentation velocity studies weie done in a Beckman model E analytical ultracentrifuge equipped with an ultraviolet scanner. Protein was prepared for these studies by dialysis against 2 liters of buffer containing 0.01 M Tris (pH 8.5), 0.1 M NaCl, and 0.01 M P-mercaptoethanol.
These studies were cariied out at a speed of 60,000 rpm.

Enzyme Reduction and Carboxymethylation
Heat Step-The pH of the supernatant fraction was increased to 7.0 with 1 M Tris base. This fraction was then distributed in 500.ml fractions into l-liter Erlenmeyer flasks. A temperature of 61" was attained in 5 min by immersion of the flask, with vigorous agitation, into a water bath at 70". The temperature of the contents of the flask was maintained at 61" by agitation in a 61" water bath for 8 min. Rapid cooling to 5" was achieved by immersion of the flask in an ice-salt bath. A supernatant fraction was obtained after centrifugation at 27,000 X g for 50 min. The heat step generally gave a a-fold purification with 75 to 85% recovery of the input activity.
Reduction and carboxymethylation of the enzyme were carried out by the method of Craven, Steers, and Anfinsen (17).

Trypfic Digestion and Pepfide Mapping
Lyophylized carboxymethylated enzyme was dissolved to 1 mg per ml in 0.2 N ammonium bicarbonate.
Three aliquots of trypsin were added to a final enzyme to substrate ratio of 1:40 (w/w), and the digestion mixture was incubated at 37" for 5 hours. Following lyophilization of the tryptic digest, peptide maps were prepared on Whatman No. 3MM chromatography paper by the method of Katz, Dreyer, and Anfinsen (18). The chromatography in l-butanol-acetic acid-water (4 : 1: 5) was performed without prior equilibration of the paper. Electrophoresis in pH 3.6 pyi,idine acetate buffer was carried out for 110 min at 2000 volts. Peptides were stained with the cadmium-ninhydrin reagent of Dreyer and Bynum (19).

Ammonium
Sulfate Fracfionation-Ammonium sulfate, 22.6 g/100 ml of supernatant fraction, was added slowly with stirring (4). The pH was maintained at 7.0 by addition of 1 M Tris base. The ammonium sulfate suspension was stirred for 30 min, and then was centrifuged at 27,000 x g for 20 min. Twelve grams of ammonium sulfate per 100 ml of original volume were added to the supernatant solution as described above. Following centrifugation at 27,000 x g for 20 min, the precipitate was redissolved in 40 to 60 ml of Buffer A. This material was placed in a dialysis bag that had been boiled for 5 min in lo-* M EDTA. Dialysis against 4 liters of Buffer A at pH 8.0 was performed for 3 hours.

Protein Determination
Sephadex G-150 Chromatography-The dialyzed ammonium sulfate fraction was applied to a column of Sephadex G-150, 229 x 5 cm, 4.3 liters, previously equilibrated with standard buffer. The enzyme was eluted with this buffer at a rate of 40 to 50 ml per hour. After the passage of approximately 1800 ml of buffer, enzymatic activity appeared just behind the excluded material.
Fractions of 20 ml were collected, and those containing enzyme with the highest specific activity were combined. Protein concentration, during the purification procedure, was estimated by the method of Lowry et al. (20) with insulin as a standard.
The protein determination of fractions prior to the Sephadex G-150 chromatography step was performed on material that had been passed through a Sephadex G-50 column (13). The nitrogen content of the purified enzyme was assayed by the micro-Kjeldahl technique. Quadruplicate analyses were conducted by the method of Ma

DEAE-Sephadex
A-50 Chromatography-The volume of the combined fractions from t,he Sephadex G-150 step was doubled with cold-distilled water.
To this solution, 2.8 mmoles of p-mercaptoethanol per 100 ml were added and the pH was increased to 8.6 with 1 M Tris base. DEAE-Sephadex A-50, equilibrated with standard buffer at pH 8.6 and 0.05 M NaCl, was used to fill a column, 36 x 2.5 cm, of 150-ml bed volume.
Following the application of the enzyme solution to the column, 1 to 2 column volumes of the equilibration buffer were passed through the column.
The protein was eluted with a 1400.ml continuous linear gradient of NaCl (0.05 to 0.5 M) in the equilibration buffer. The flow rate of the column was 0.5 ml per min, and fractions of  (14). were concentrated 3-fold in an Amicon ultrafiltration cell equipped with a UM-1 filter.' The concentrated enzyme solu-Electrophoresis of 275 pg of protein in polyacrylamide gels contion was then made 3 M in NaCl* and 0.01 M in dithiothreitol taining SDS showed one major band and one minor band (see and stored at -20". The specific activity of the purified enzyme below).
stored as indicated decreased 50% over a period of 2 weeks.
A summary of this procedure is presented in Table I. Similar Puri$catwn of Feedback-resistant Enzyme results have been obtained on three other occasions. The final Phosphoribosyltransferase was also purified from the mutant, concentrated enzyme has a specific activity of 8800 units per mg hisGIl hisIF (5), in which the feedback inhibition site of of protein based on the coupled G-70 assay and a protein deter-the enzyme is insensitive to histidine. The purification promination by the method of Lowry et al. (20). A specific activity cedure was identical with that described for the wild type enzyme of 3000 units per mg of protein was obtained when the enzyme except for the following differences: (a) the histidine concentrawas assayed by the pyrophosphatase assay. Quadruplicate ni-tion in the standard buffer was increased to 0.01 M, and (b) the trogen analyses were performed on the purified enzyme following cells were disrupted by two passages through a Gaulin Laboraextensive dialysis against distilled water. The nitrogen content tory homogenizer9 at 10,000 p.s.i. Preliminary experiments had and the percentage weight nitrogen in the enzyme (18%) were shown that 0.01 M histidine was required to stabilize the mutant used to calculate a protein concentration. On the basis of this enzyme for the heat step. The mutant enzyme behaved in a value and with the coupled G-70 assay, the specific activity of the similar manner to the wild type enzyme throughout the purificaenzyme was 6600 units per mg of protein. Velocity-At a protein concentration of 1.4 mg Electrophoresis of 50 pg of purified feedback-resistant enzyme in per ml, a single symmetrical peak with an ~20,~ of 8.94 S was ob-polyacrylamide gels containing urea showed a pattern identical served. This value is in good agreement with the value of 8.83 S with that observed with the wild type enzyme. One major and reported previously (4).
one minor band were noted on SDS gel electrophoreais. During Disc Gel Electrophoresis-Electrophoresis of 50 pg of protein sedimentation velocity studies at a protein concentration of 1.4 on polyacrylamide gels containing urea (14) revealed one dark mg per ml, a single symmetrical peak with an .sz~,~ of 9.16 S was staining band and two minor bands (Fig. 1). At 100 hg of pro-observed. Insufficient analyses were carried out to determine tein two additional faint bands appeared. Although these minor whether or not this value is significantly different from 8.94 S. bands may represent contaminants, the possibility that they are isozymes or aggregates has not, been ruled out. On electrophore-Properties of Puri$ed Wild Type Phosphoribosyltransferase sis in polyacrylamide gels without urea multiple bands appeared.

Kinetics of Forward Reaction-A
constant magnesium (MgC12) to ATP ratio of 2: 1 (except Fig. 4, A and B)   The apparent Km for PP-ribose-P (with the standard ATP concentration) was obtained from a Lineweaver-Burk plot (22) and is 5.6 X 1OW M (Fig. 2). The apparent li, for ATP (with the standard PP-ribose-P concentration) is 4.3 x 10m4 M (Fig. 3). These Lineweaver-Burk plots do not deviate from linearity. In order to derive concentration-independent Michaelis constants, secondary plots of the ordinate intercepts against l/[fixed substrate] were drawn (Figs. 2 and 3, insets). The concentrationindependent Michaelis constant for ATP is 4.2 x 10d4 M (Fig. 2,  inset), while the concentration-independent Michaelis constant for PP-ribose-P is 1.0 x lop4 M (Fig. 3, inset). When the magnesium concentration was maintained at 10 mM, the apparent Km for PP-ribose-P (standard ATP concentration) was 5.1 X lOA M (Fig. 4A) and that for ATP (standard PP-ribose-P concentration) was 4.8 X 10m4 M (Fig. 4B). Double reciprocal plots of initial velocity against PP-ribose-P concentration at different fixed concentrations of ATP resulted in a family of lines that intersect to the left of the ordinate above the abscissa (Fig. 2). The same pattern was seen when ATP concentration was varied as a function of several fixed concentrations of PP-ribose-P (Fig. 3). The nonparallel pattern of the initial velocity plots suggests that under these conditions the predominant mechanism of the forward reaction is one involving a ternary complex (23).
Reversibility of Reaction-The reversibility of the phosphoribosyltransferase reaction, i.e. formation of PP-ribose-P and ATP from 1%ATP and PPi in the presence of Mg+f, has been shown (1). Table II depicts the absolute requirement of this reaction for enzyme, both substrates, and magnesium.
Substitution of PR-AMP for PR-ATP yielded no reaction. Interestingly, the reverse reaction was inhibited by histidine; 50y0 inhibition was observed at 4 X 10e4 M histidine.
Feedback Inhibition of PurQied Phosphoribosyltransferase L-Hi&dine inhibition- Table  I  50%. Previously reported preparations of the purified enzyme were inhibited 5070 by 6 X 10-S M and 9 X 10V5 M L-histidine (4).
When the PP-ribose-P concentration was varied (Fig. 4A) The pH of the assay mixture was measured at the end of each assay. and I(, (24). Inhibition by L-histidine was noncompetitive (24) when the ATE' concentration was varied (Fig. 4B). Fiv. 5 illustrates that a sigmoid curve was obtained when the h activity of the forward reaction was followed as a function of Lhistidine concentration.
A secondary plot of these data (Fig. 6), with the modified Hill equation (27), yielded a linear plot with a slope (n) of 1.6. The fact that the slope is greater than 1 and less than 2 suggests the presence of at least two interacting binding sites on the enzyme for histidine.
The value of n varied with the age of the enzyme and approached 1 in older preparations.
Thimoleakmine Inhibition-Moyed and Friedman (28) and Moyed (29) initially showed that thiazolealanine, a histidine analogue, mimics L-histidine in its inhibition of the BBM IIIsynthesizing system of Escherichia coli. Ames et al. (1) and Martin (3) subsequently demonstrated that the inhibition observed was due to inhibition of the phosphoribosyltransferase. JYhen the rate of the forward reaction was followed as a function of thiazolealanine concentration (with respect to the L isomer), a sigmoid curve was observed (Fig. 7). Fifty per cent inhibition occurred at 4 X 10W4 M thiazolealanine. Use of the modified Hill equation yielded a straight line with a slope (n) of 1.8 (Fig. 8). This value of n is also consistent with the presence of at least two interacting binding sites on the enzyme for thiazolealanine.
The value of 12 was 1 in older preparations of the enzyme.
L-H&k&e and Thiazolealanine Inhibition as Function of pH- Fig. 9 shows that feedback inhibition by L-histidine and thiazolealanine (with respect to the L isomer) is pH-dependent with  maximal inhibition by each effector occurring in the physiological pH range. The curve for inhibition by each molecule is similar to that expected for the titration curve of a single group with an approximate pK of 8.7. In a similar experiment with partially purified enzyme, Martin (3) reported a value of 9.2. The pK of the a-amino group of L-histidine is 9.2 (30). The difference between the pK obtained in this experiment and the pK of L-histidine suggests that t,he approximate pK of 8.7 represents the pK of the interaction of L-histidine with a group (groups) on the enzyme.

Comparison of Wild Type and Feedback-resistant Phosphoribosyltransferases
Subunit Molecular Weight in SDS Gels-The mobility of both enzymes in polyacrylamide gels containing SDS was measured with proteins of known molecular weights as standards (15). The mobilities of the wild type and feedback-resistant enzymes were identical, 0.66 (Fig. 10). These mobilities correspond to an approximate molecular weight of 34,000 for the subunits of the wild type and feedback-resistant enzymes. A minor band of lower mobility was occasionally observed in both preparations.
Tryptic Peptide Maps-Peptide maps of the wild type enzyme were reproducible and revealed 34 to 36 ninhydrin-positive spots plus some core material at the origin.
The peptide map of the feedback-resistant enzyme was indistinguishable from that of the wild type enzyme.
Sedimentation Velocity Studies-As reported above, the wild type enzyme had an szo,W of 8.94, while the feedback-resistant enzyme had an szo ,W of 9.16. These values agree to within 3%.
Difference Spectra-A striking change was observed in the difference spectrum at 280 mp of the wild type enzyme upon the addition of L-histidine (Fig. 11). The absorbance at 280 amino acid residues are being exposed. However, no precipitate was observed in the cuvettes and the spectra do not appear to be consistent with the formation of a precipitate (absorption is not proportional to 1/X4) although anomalous absorption due to precipitation cannot be excluded. The difference spectrum of the feedback-resistant enzyme remained constant (Fig. 12) following the addition of histidine. Whatever may be the basis for the a,nomalous absorption of the wild type enzyme upon the addition of histidine, these data suggest that the interaction of L-histidine with the enzyme leads to a conformational change which may be necessary for feedback inhibition of the enzyme. DISCUSSION The purified phosphoribosyltransferase was used to study the kinetic properties of the forward reaction.
Measurements of velocity as a function of substrate concentration showed normal Michaelis-Menten kinetics for PP-ribose-P and ATP even in the presence of histidine.
These results suggest that neither PP-ribose-P nor ATP undergoes homotropic interactions.
Initial velocity analysis, under optimal assay conditions, resulted in a family of intersecting lines on double reciprocal plots for both PP-ribose-P and ATP. Such kinet'ic patterns imply that under these conditions the predominant reaction sequence for PR-ATP synthesis is one involving a ternary complex of PP-ribose-P, ATP, and enzyme (23). In contrast to these results, it was earlier shown (3) that the partially purified enzyme catalyzes an exchange between 32PPi and PP-ribose-P in the absence of ATP and between 14C-ATP and PR-ATP in the absence of PP-ribose-P.
Recently Bell and Koshland (34), using purified phosphoribosyltransferase obtained by a modification of the procedure described in this paper, have isolated a phosphoribosyl-enzyme intermediate.
The existence of a covalent enzyme-substrate complex suggests that under other circumstances PR-ATP synthesis may proceed via a "ping-pang" mechanism in which release of the first product precedes addition of the second substrate.
The factors which cause the predominance of one reaction sequence over an alternative one are not known.
A general model (25) proposed to explain the properties of regulatory enzymes distinguished two types of regulatory systems (K and V) on the basis of kinetic properties.
The kinetic properties of the phosphoribosyltransferase are closest to those of a negative Ti system. Neither substrate exhibits homotropic interactions, while the inhibitors, histidine and its analogue, thiazolealanine, both exhibit homotropic interactions.
Alternatively, the kinetic properties of the enzyme could be accounted for by other models such as the one described by Koshland,Nemethy,and Filmer (35).
A guiding interest in this study has been the mechanism of feedback inhibition by histidine.
Histidine was observed to be an uncompetitive inhibitor of the wild type phosphoribosyltransferase with respect to PP-ribose-P and gave noncompetitive inhibition with respect to ATP. It therefore seems unlikely that histidine inhibits by merely binding to the active site.
The sigmoid curves for histidine and thiazolealanine inhibition are consistent with the cooperative binding of these effecters to the enzyme. When these curves are transformed into straight lines by use of the modified Hill equation, the slopes (n) are greater than 1. Since n is an interaction coefficient (25,26), values of n greater than 1 indicate the presence of at least two (possibly identical) interacting binding sites on the enzyme for these effecters.
The similarity of the pH dependence of histidine and thiazolealanine inhibition suggests that the two inhibitors are acting at similar if not identica,l binding sites on the enzyme. Sigmoid curves for histidine inhibition have also been reported (36) for partially purified E. coli phosphoribosyltransferase. Evidence for the existence of more than one type of binding site comes from two quarters.
The feedback-resistant mutant, hisGil hisIFlS5, requires a higher concentration of histidine to protect it during the heat step than does the wild type enzyme. This suggests that histidine is still capable of binding to the feedback-resistant enzyme, thereby stabilizing it to thermal inactivation.
Second, Martin (3) has shown that 3H-I.-histidine binds equally well to histidine-sensitive and to histidine-insensitive enzyme.
It cannot be assumed that the histidine binding site associated with stabilization or binding in these studies is necessarily required for feedback inhibition, although such an explanation is attractive.
These observations, together with the results of the Hill plots, can be explained by postulating that histidine inhibition depends upon the binding of histidine to at least two interacting binding sites. Thus, whereas binding of histidine is necessary for feedback inhibition, it is not sufficient unless binding to at least a second site occurs, and there is some interaction between the two sites.
The feedback-resistant enzyme was purified in order to find any difference in the response of this enzyme to histidine.
The purified wild type and feedback-resistant enzymes were found to have similar sedimentation coefficients, subunit molecular weights, and tryptic peptide maps. The two enzymes, therefore, are quite similar in their physical and chemical properties.
Since the feedback-resistant mutation lies in the operator-proximal third of the G gene (37), it is likely that the mutation is a missense, rather than a nonsense or frameshift, mutation.
Mutations of the latter two types would produce a truncated protein (13,37,39). Therefore, the two enzymes probably differ only by the substitution of a single amino acid.
The most striking divergence between the wild type and feedback-resistant enzymes was in the response of their individual difference spectra to L-hietidine.
The difference spectrum of the wild type enzyme was still increasing in absorbance at 280 rnp 170 min following the addition of histidine.
In contrast, no change in the difference spectrum of the feedback-resistant enzyme occurred upon the addition of histidine.
The fact that the difference spectrum of the feedback-resistant enzyme remained constant makes it unlikely that there is a trivial explanation for the change in the difference spectrum of the wild type enzyme. These findings are consistent with t'he occurrence of a conformational change in the enzyme when histidine interacts with the intact feedback inhibition sites possibly followed by precipitabion of the wild type enzyme.
Presumably in the feedbackresistant enzyme, these sites are altered or interaction with histidine does not produce the conformational change (or both). It has recently been shown that binding of histidine to the wild type phosphoribosyltransferase is accompanied by a conformational change in the enzyme in which 12 previously exposed tyrosyl residues become buried.'O Finally, on the basis of a molecular weight of 215,000 for the native enzyme, a subunit molecular weight of 35,000, and an amino acid analysis, Voll et al. (4) calculated that, if the phosphoribosyltransferase were composed of six identical subunits of 35,000 molecular weight, tryptic digests of the enzyme would 'ODrs. F. Blasi, S. M. Aloj, and R. F. Goldberger, personal communication.
These authors reported that a preliminary tryptic peptide map of the digested protein gave 34 to 42 tryptic fragments plus a small amount of core material.
The tryptic peptide maps of the wild type phosphoribosyltransferase reported in this paper were reproducible and gave 34 to 36 ninhydrin-positive spots plus a small amount of core material. A subunit molecular weight of approximately 34,000 was obtained in SDS polyacrylamide gel electrophoresia.
These results in conjunction with the earlier observations of Voll et al. (4) indicate that the wild type and feedback-resistant phosphoribosyltransferases are composed of six similar subunits.