Mechanism of Porphobilinogen Synthase POSSIBLE ROLE OF ESSENTIAL THIOL GROUPS*

The inactivation of beef liver porphobilinogen synthase by the alkylating agents iodoacetate and iodoacetamide has been investigated. The synthase can be inactivated by these reagents only if the enzyme has been activated initially by sulfhydryl reducing agents. Inactivation by both iodoacetate and iodoacetamide was the result of modification of cysteinyl residues. A plot of the incorporation of W-labeled alkylating reagents uersus activity extrapolated to the modification of approximately one fast reacting cysteinyl residue per monomer (M, = 35,000) as responsible for the loss of activity by either iodoacetate or iodoacetamide; the pH of half-maximum alkylation, with both reagents, was about 5.2 indicating that each reagent alkylates a cysteinyl residue of similar high reactivity. The substrate analogue, levulinic acid, a competitive inhibitor, partially protected the enzyme from inactivation by iodoacetate, suggesting that this alkylation was active site directed. In contrast, levulinic acid potentiates the inactivation caused by iodoacetamide. The product of the reaction, porphobilinogen, however, partially protected the enzyme from inactivation by both iodoacetate and iodoacetamide. Equilibrium

The inactivation of beef liver porphobilinogen synthase by the alkylating agents iodoacetate and iodoacetamide has been investigated.
The synthase can be inactivated by these reagents only if the enzyme has been activated initially by sulfhydryl reducing agents. Inactivation by both iodoacetate and iodoacetamide was the result of modification of cysteinyl residues. A plot of the incorporation of W-labeled alkylating reagents uersus activity extrapolated to the modification of approximately one fast reacting cysteinyl residue per monomer (M, = 35,000) as responsible for the loss of activity by either iodoacetate or iodoacetamide; the pH of half-maximum alkylation, with both reagents, was about 5.2 indicating that each reagent alkylates a cysteinyl residue of similar high reactivity.
The substrate analogue, levulinic acid, a competitive inhibitor, partially protected the enzyme from inactivation by iodoacetate, suggesting that this alkylation was active site directed.
In contrast, levulinic acid potentiates the inactivation caused by iodoacetamide. The product of the reaction, porphobilinogen, however, partially protected the enzyme from inactivation by both iodoacetate and iodoacetamide.
Equilibrium dialysis studies of [Wllevulinic acid binding to the enzyme have shown that native enzyme (thiol-activated or air-oxidized) and the alkylated enzyme derivatives bind the same maximum number of moles of levulinic acid/ mol of enzyme (2/285,000 daltons). However, the binding affinities of the native thiol-activated enzyme and the iodoacetamide-inactivated enzyme were 3-to 5-fold greater than those of the air-oxidized and iodoacetate-inactivated samples of enzyme.
Peptide ated a separate essential cysteinyl residue in the primary sequence. A mechanism is proposed in which cysteinyl residues may participate in the acid/base catalysis required for the enzymic protonationldeprotonation sequences in the synthesis of porphobilinogen. TPCK, L-1-tosylamido-2-phenylethylchloromethvl ketone. ' G. F. Barnard and D. Shemin, (1976 continuously, and aliquots of 25 ~1 were counted for '"C radioactivity.
The columns were loaded with either 12 mg of tryptic peptides in 100 ~1 of 50 rnM NH,HCO, or with fractions eluted from the PA-35 column which had been lyophilized and resuspended in 50 mM NH,HCO,.

Inactivation
by IAm and IAc -PBG synthase requires thiol reducing agents for enzymic activity. Air-oxidized samples of PBG synthase have at most 0.002% of the specific activity of the enzyme activated with dithiothreitol.
Consistent with this thiol requirement for activity is the observation that alkylation by IAm or IAc (8,9) inactivates PBG synthase (Fig. 1, A and B). The extent of inactivation is dependent upon the concentration of the alkylating reagent but, as can be seen from the semilog plots, the rate of the irreversible alkylation deviates from linearity.
This can be attributed to the simultaneous alkylation of the dithiothreitol continuously reducing the effective concentration of the alkylating reagents. Similar complications have previously been observed (36). Furthermore, as shown in Fig. 2, alkylation does not occur unless the enzyme is first reduced with dithiothreitol. Since, as shown below, the alkylation occurs on cysteinyl residues, the reduction by thiol reducing agents not only produces -SH groups essential for catalysis, but also exposes these groups for alkylation.
Identification of Amino Acid Residue Modified by Alkylation -Since the thiol activation of PBG synthase determines the susceptibility of the enzyme to inactivation by alkylation (Fig. 21, it was considered that cysteinyl residues may be the site(s) of modification.
However, since under various conditions it is known that IAm and IAc can modify other residues such as histidine (37-401, methionine (41, 421, and glutamate (431, it was important to identify the residue(s) that were alkylated. As a preliminary indication of the nature of the residues that were alkylated, inactivation by IAm and IAc was carried out after DTNB had been added to a thiolactivated sample of PBG synthase (as described under "Materials and Methods"). The titration of the enzyme with DTNB fully protected the enzyme against inactivation by either IAm or IAc (Table I). This result strongly suggested that cysteinyl residues were alkylated.
To identify the modified residues conclusively, the enzyme was inactivated with [W]IAm or [W]IAc and hydrolyzed with HCl. It can be concluded from Fig. 3, A, B, and C, that both IAm and IAc alkylate only cysteinyl residues. Identical conclusions were obtained from results in all three systems  were removed into 0.95 ml of standard assay medium at 2". After 10 min of preincubation at 37", the assay was started by the addition of &aminolevulinic acid (5.0 mM) and incubated for 15 min. with samples of hydrolysates of the enzyme alkylated either with IAc or IAm. The alkylation of a cysteinyl residue of the synthase by IAc was recently reported (9). Number of Cysteinyl Residues Modified -To determine the number of residues modified that lead to complete loss of enzymic activity, the degree of inactivation was measured as a function of the extent of alkylation. The extent of alkylation was determined by measuring the 14C incorporation following alkylation with [l-W]IAm or [l-W]IAc. The extent of the inactivation was varied either by using different concentrations of the reagents for a fixed time at constant pH or by alkylating with a fixed concentration of the reagent at different pH values. It can be seen from Fig. 4, A and B, that in each case the initial slope extrapolates to a value of 0.8 to 1.0 mol of the alkylating agent incorporated/monomer of 35,000. This implies that primarily one essential residue was modified in each case. (The slope deviated from linearity as inactivation exceeded about 80%.) pH Dependence of Alkylation -The pH profile for alkylation was determined as a measure of the reactivity of the cysteinyl residues toward IAm and IAc. The reactivity was determined simultaneously by following both the loss of enzymic activity and the incorporation of 14C radioactivity at various pH values. The results given in Fig. 5, in which the residual activity is plotted against pH, demonstrate in both experiments that the inflection point occurred at pH 5.2. The experimental data obtained with IAm were in good agreement with a theoretical titration curve for a single ionizable group. The results of experiments with IAc were somewhat anomalous; at lower pH values some inactivation still occurred. This implied that other factors are interfering with the titration of a single ionizable group. However, these experiments illustrate that both reagents alkylate cysteinyl residues of high reactivity. Similar results were obtained in plots of pH versus 14C radioactivity incorporated. Alkylation in Presence of Levulinic Acid or Porphobilinogen -A criterion for establishing whether a particular modification is active site directed is to determine whether the substrate or product protects against inactivation.
It can be seen that the presence of PBG protected against the inactivation by either IAm or IAc (Fig. 6) active site. The assay procedure described in the legend to Fig. 6 was necessary because of the inhibition of the enzymic activity by PBG (44).
Levulinic acid, a substrate analogue, is a competitive inhibitor of PBG synthase with a Kj of 1.2 x 10m4 M (5) (0)).
was used to avoid product formation and yet provide active site protection. The presence of levulinic acid in alkylation experiments uncovered a difference between the characteristics of inactivation by IAm and IAc. It can be seen from Fig.  7A that, whereas levulinic acid protected against the inactivation by IAc, it augmented the inactivation by IAm (Fig.  '7B). This difference raised the possibility that the two reagents alkylate different cysteinyl residues of the primary sequence.
Although the alkylation by IAm was not protected by levulinic acid, the finding of protection by PBG suggested that the cysteinyl residue modified by IAm is near or at the active site which consists of two &aminolevulinic acid binding sites, A and P (Scheme 1).
[W]Levulinic Acid Equilibrium Binding to PBG Synthase -Since the alkylations of the enzyme by IAc and IAm were affected differently by levulinic acid, we investigated the binding characteristics of these alkylated derivatives of the enzyme and compared the results with those obtained with air-oxidized and thiol-reduced native enzyme. The binding of [Wllevulinic acid to both native and modified samples of PBG synthase was studied by equilibrium dialysis at 2". The modified PBG synthase samples used in these experiments had lost at least 99.7% of their original activity by alkylation. The results shown in Fig. 8 expressed as Scatchard plots (45) where d is the number of moles of ["Qlevulinic acid bound/m01 of 285,000. It can be seen in Fig. 8 and Table II that, whereas both alkylated samples of PBG synthase bound a maximum of 2 mol of levulinic acid the affinity of the enzyme alkylated with IAm for levulinic acid was approximately 3-to 4-fold greater than that of the enzyme alkylated with IAc. These results were very similar to those obtained with native thiol-reduced enzyme and native air-oxidized PBG synthase (Fig. 8 and Table II).4 It can be seen that the binding properties of the enzyme alkylated with IAm are similar to the native active enzyme, whereas the enzyme alkylated with IAc has similar properties to the air-oxidized preparation.
It is, however,interesting to note that the inactivated samples of the enzyme bind the same maximum number  Fig. 7. Inactivation by IAm is not protected by levulinic acid nor is the binding of levulinic acid altered after IAm alkylation; similarly, inactivation by L4c is partially protected by levulinic acid and the affinity for levulinic acid binding is decreased after alkylation with IAc.
Sites of Cysteine Alkylation-The data obtained from the three experimental approaches described below imply that the cysteinyl residues modified by IAm and IAc are in distinct regions of the primary sequence of PBG synthase. Samples of PBG synthase which were labeled with [14ClIAm or [WlIAc in a manner to insure incorporation of less than 1 mob monomer were digested with trypsin as described under "Materials and Methods" and the radioactive peptides were identified after (a) fingerprint mapping at pH 2.5 and pH 6.5; (b) molecular sieve chromatography; and (c) cation exchange chromatography.
In all these procedures the chromatographic behavior of the IAc-derived W-labeled peptide was different from the IAm-derived major '%-labeled peptides. These differences could conceivably be attributed, in part, to a charge difference between the amide and carboxylate groups on the alkylated peptides. However, it can be seen from the fingerprint mapping carried out at pH 2.5 (Fig. 9, A and B) and at pH 6.5 (Fig. 9, C and D) that the different mobilities of the labeled peptides occurred by chromatography rather than by electrophoresis, and two radioactive peptides were obtained on co-chromatography ( Fig. 9 E). The elution profiles obtained from similarly prepared tryptic digests of alkylated enzyme on cation exchange chromatography were also different ( Fig.  10, A and B). Since the IAm peptide was eluted in a position corresponding to a more acidic peptide than the position of the IAc-labeled peptide, the elution position of the peptides labeled with IAm or IAc cannot be explained in terms of the difference in charge if the same peptide was alkylated.
Furthermore, the profiles of radioactive peptides obtained by molecular sieve chromatography of tryptic digests of [W]IAm and [14C]IAc-labeled enzyme were again different (Fig. 11, A and B). The only significant radioactive peak from the IAc-labeled enzyme centered around Fraction 54 (Fractions 52 to 55), whereas the IAm-labeled enzyme produced three major radioactive fractions. Redigestion with trypsin of  the latter larger pieces produced more of the radioactive peptide eluting in Fractions 79 to 84. The relative position of the fractions of the IAc-labeled peptide (Fractions 52 to 55) and the major IAm-labeled peptide (Fractions 79 to 84) were maintained on co-chromatography after redigestion with trypsin.
whereas the major labeled peptide obtained from the enzyme alkylated with IAm was found to be a dipeptide consisting of Cys(Cm) and Tyr. DISCUSSION These data demonstrate that the peptides isolated after alkylation by IAm and IAc and tryptic digestion are indeed different. This conclusion was further supported by amino acid analysis of radioactive peptides isolated and purified by a combination of the above procedures. The peptide labeled with IAc was found to be an octapeptide with the following composition; Cys(Cm), Glx, Pro, Gly, Ala, Val, Tyr, Arg, The results of this communication demonstrate that IAc and IAm each irreversibly inactivate PBG synthase by alkylation of highly reactive essential cysteinyl residue(s) probably located at the active site. Furthermore, we have shown that the inactivation by alkylation only occurs if the enzyme is initially reduced with thiol compounds; the state of cysteine reduction of PBG synthase is all important for activity. It has been claimed that thiol activation generates two cysteinyl  With the demonstration that the -SH group(s) are probably at the active site, it seems reasonable to expect that these group(s) would be more reactive than those in normal cysteinyl residues. As a measure of the reactivity, the pH dependence of alkylation by IAm and IAc was investigated. With both reagents the inflection point of half-maximum reaction was at pH 5.2. This is 3 to 4 orders of magnitude lower than the pK, of nonactivated cysteinyl residues (55-57). The pK, values of essential cysteinyl residues in other enzymes have been estimated to vary between 5.0 and 9.2 (16,20,53,55,(58)(59)(60). The lower pK, values imply the stabilization of the thiolate anion. This may be achieved by thiol-base pair formation [---Se . . H . @B] and has been suggested as a means of lowering the pK, of the essential cysteinyl residues in several other enzymes (53,54,58,61,62) and could play a role in the mechanism of PBG formation by PBG synthase. Since the alkylation rates were not measured above pH 7.5 because of competing alkylation of dithiothreitol, it is possible that the measured pK, values of PBG synthase are only the acid limbs of a bell-shaped pH-dependent profile as observed with papain (63, 64) and streptococcal proteinase (19).
In view of the different effects of levulinic acid on alkylation (protection against IAc, potentiation toward IAm inactivation) the different binding characteristics of the alkylated enzymes (reduced affinity after IAc alkylation, unaltered affinity after IAm alkylation), the similar protection by PBG toward alkylation by both reagents, the demonstration that different tryptic peptides are labeled by each reagent (by chromatographic and amino acid analysis), it would appear that IAc and IAm alkylate cysteinyl residues in different positions of the primary structure.
On examination of the composition of the octapeptide obtained from IAc alkylation and the dipeptide obtained from IAm alkylation, a rational explanation that the dipeptide is a subset of the octapeptide can be advanced if one assumed that the sequence at the carboxyl end of the octapeptide is -Arg-Cys(Cm)-Tyr, and that this peptide is the COOH-terminal end of the protein. This is reasonable for it is known that an acidic amino acid on the COOH-terminal side of arginine or lysine inhibits tryptic cleavage5 (65). The protein alkylated with IAm, therefore, would more readily be hydrolyzed by trypsin between the arginine and carboxyamidomethylcysteine. If this were so, the dipeptide would be a subset of the octapeptide and both IAm and IAc alkylate the same cysteinyl residue. However, on sequencing the octapeptide the NH,-terminal amino acids were Cys(Cm)-Tyr-Gln, thus eliminating this possibility. Preliminary studies on the sequence of the dipeptide by dansylation indicated that the dipeptide is Cys(Cam)-Tyr, for only the 0-dansyltyrosine was obtained.6 (It appears that the NH,-terminal Cys(Cam) had cyclized) (66,67). If the dipeptide has the same sequence as these two residues in the octapeptide the possibility still exists that the dipeptide is a subset of the octapeptide. This peptide could arise by an anomalous cleavage by the trypsin-5 This possibility was raised in a discussion with Dr. George R.

Stark of Stanford
University. @ We wish to thank Thomas Ott for this experiment. TPCK at tyrosine. However, there is no precedent that can be cited in which this anomalous cleavage is influenced by the nature of the modified cysteinyl residue NH,-terminal to the tyrosine residue.
Furthermore, if the dipeptide were a subset of the octapeptide obtained from the IAc-labeled enzyme, it could be expected that the larger peptides obtained on tryptic digestion of the IAm-labeled enzyme would have a composition related to the octapeptide.
Preliminary experiments in which an IAm-labeled enzyme preparation was digested with trypsin for a limited time yielded some dipeptide and a major large peptide (about 20 residues) which contained lysine but no arginine, the latter being the COOH-terminal residue of the octapeptide. This finding suggests that the dipeptide is not a subset of the octapeptide.
Although the occurrence of two Cys-Tyr sequences in one polypeptide may be considered to be rare, it is not without precedent occurring in bovine prothrombin, porcine phospholipase A2, and in several neurotoxins.' In view, however, of some possible misgivings, the final decision regarding the alkylation of different cysteinyl residues by IAm and IAc must await further exploration.
Experiments are in progress to elucidate this point (see "Note Added in Proof").
The mechanism of PBG formation as proposed by Nandi and Shemin (5) depicts an aldol condensation between one molecule of b-aminolevulinic acid, which is in Schiff base linkage with the A site of the catalytic site, with a second molecule of &aminolevulinic acid at the P site. Inspection of this mechanism suggests the requirement for several enzymic protonation/deprotonation sequences catalyzed by particular residues at the active site.
The finding that cysteinyl residue(s) at the active site are of high "reactivity" suggests that these essential residues may participate in the necessary acid/base catalysis by the enzyme. A similar acid/base role has been proposed for a cysteinyl residue in proline racemase (23) and histidine decarboxylase (55). In order to participate throughout the catalysis as depicted in Fig. 12, each group (B-or BH) must alternate its acid/base status. A convenient means of achieving this is to interpose an imidazole bridge (or even a water molecule*) for regeneration of the required state. Consistent with this suggestion is the observation that a histidinyl residue is essential for PBG synthase catalysis (6).
The mechanism in Fig. 12 involves the initial deprotonation, by B-(possibly a thiolate anion) of the S-aminolevulinic acid molecule at the A site and the protonation of the carbonyl group by BH (possibly thiol) at the P site (Fig. 12, Step 1). The starting B-and BH status is regenerated by mediation of the imidazole bridge (Step 2). This is followed by a second protonationldeprotonation sequence with the elimination of water (Step 3). The subsequent transamination steps together with the mediation of the imidazole bridge system complete PBG formation and regenerate the initial state of the enzyme (Steps 4 to 7).
The mechanism in Fig. 12 is modified from that of Nandi and Shemin to incorporate the observation that the deprotonation of the immediate precursor of free PBG (Step 6) is stereospecific and, therefore, enzyme-catalyzed (68, 69). The important feature of our mechanism is the bridged system for