Structural and Functional Roles of the Cysteine Residues in the a Subunit of the Escher-k?& coli Tryptophan Synthetase

the LY subunit purification, and to Clysses Smalls and Joanne Chcst’er of the Department’s Service Facilit)y for ultracentrifuge runs and amino acid analyses.

erichia co2i tryptophan synthetase have been cross-linked by use of the bifunctional reagent N,N'-bis(maleimidomethyl) ether (BME). The sites of reaction were identif%ed from the amino acid compositions of radioactive peptides purified from tryptic and tryptic-chymotryptic digests of Q! subunit which had reacted with 3H-BME to the extent of 1 mole of cross-linking reagent per mole of enzyme.
Hydrodynamic studies showed that no significant dimer formation had occurred. The electrophoretic and immunochemical properties of the enzyme are identical with those of untreated control preparations. Although the intramolecularly cross-linked monomers are apparently homogeneous by these criteria, BME-treated a! subunits are functionally heterogeneous in that 70% of the independent indoleglycerol phosphate activity of the o( subunit has been lost after reaction of 1 mole of BME with 1 mole of a! subunit. In addition, 70% of the protein in such preparations cannot form a functional c@&yptophan synthetase complex with the PZ subunit, as measured indirectly by enzymatic activity studies and directly in studies on C&Z complex formation in sucrose density gradient centrifugation experiments.
Two possible explanations for the observed functional heterogeneity of BME cysteine 80 to cysteine 117 crosslinked a! subunit preparations are advanced. These mechanisms may be interdependent. Two new asymmetric centers are introduced by the addition of the protein cystelnyl thiols across the two maleimide double bonds of the bifunctional reagent. A population of a subunits cross-linked be- tween these 2 cysteinyl residues might contain as many as four different conformational forms of BME cross-link. The sulfhydryl group region of the enzyme is known from previous studies to be closely related to the active site of the subunit in the indoleglycerol phosphate aldolase reaction, and to be important in the association of the a! subunit with pZ subunits to form the oLzBz-tryptophan synthetase complex. Alternatively, the order of addition of the two cysteinyl thiols to the BME molecule is not known, and it is possible that differences in either protein active site structure, or in the stereochemical course of the cross-linking reaction may be the consequence of initial reaction of BME at one or the other of the 2 cysteinyl residues. Such differences in structure between the active and inactive subunits are nonetheless not great enough to allow the physical separation of the different forms involved.
The tryptophan synthetase of Escherichiu coli is composed of two types of separable protein subunits, originally designated the A and B proteins (2). The A protein, now referred to as the OL subunit, is monomeric with a molecular weight of 29,500 (3,4). The B protein (molecular weight 99,000) (5, 6) has two identical, strongly associated polypeptide chains (5) and has been redesignated the 02 subunit (6). The structure of the fully constituted tryptophan synthetase complex is that of an a&z tetramer (3,6).
The pZ component contains pyridoxal phosphate and catalyzes tryptophan synthesis from indole and I.-serine (cserine hydrolyase (adding indole), EC 4.2.1.20)) and also catalyzes the deamination of serine and a number of other reactions (9, 10) in the absence of the o( subunit. The oc& complex catalyzes both the 1440 Cysteine.s in a Subunit of E. coli Tryptophan Xylthetaae. III \'ol. 246, X0. 5 InGPi aldolase and L-serinc Irgdrolyase reactions at higher rates than the isolated subunits , and in addition, InGP and serine are converted to t'ryptophan at a high rate without t'he detectable appearance of free indole as an intermediate (11,12). Previous studies on the effects of sulfhydryl reagents on the cy subunit (13-16) have shown that the 3 cysteinyl residues at positions 80,117, and 153 of t'he 267 amino acid sequence (17-23) are important to the tertiary titructure and catalytic function of the enzyme.
Hardman and Yanofsky studied the effects of t,he sulfhydryl reagents, iodoacetate, and NEXI, on the activity of the ac subunit in the InGP aldolase reaction (13). Both reagents inactivated the enzyme completely, yet no inactivation occurred when InGP was present at substrate levels. A maximum of 3 moles of iodoacetate reacted per mole of cy subunit, reacting equally with all three cysteinyl thiol groups.
However, a maximum of only 1 mole of NEM reacted per mole of subunit, alt,hough radioactivity from '"CKEM was distributed equally between t,he 3 cysteinyl residues.
The observed substrate protection and failure of a 2nd molecule of NEM to react (although the initial NEM molecule might react with any of the three cysteines) and the fact that the less bulky iodoacetate molecules reacted readily and completely with all 3 residues, led Hardman and Yanofsky (13) to suggest that the 01 subunit might exist in a conformation in which the sulfhydryl groups were quite close to each other spatially. The determination of the complete sequence of the (Y subunit showed that the 3 cysteinyl residues were quite distant from each other in the primary sequence of the protein molecule, at positions 80, 117, and 153 (17-23).
Thus t)he unique reactivity of these three sulfhydryl groups must reflect certain features of t,he three-dimensional structure of the (Y subunit in this portion of the protein molecule. If the cysteine side chains were extremely close to each other in the three-dimensional geometry of the active site region, one might expect intramolecular oxidation to occur. However, disulfide bonds are found neither before, nor after treatment of the OL subunit with NEM (13, 15).
To further examine the conformation of the (Y subunit in this critical region, the bifunctionnl reagent, BME, was chosen. This reagent may be considered as a bifunctional analogue of NEM, and has been used in conformational studies on hemoglobin (24,25). In it,s fully extended configuration, BILlE has the potcnt,ial to cross-link t,hiols 10 to 13 A between sulfur atoms. IIence, sulfhydryl groups too far apart from each other to be readily oxidized to disulfides, yet as close to each other as suggested by the data of IIardrnan and Yanofsky (13), might be covalently cross-linked by this reagent. This paper shows that 2 of t)he 3 cysteinyl residues can be cross-linked by BME under extremely gentle conditions. Values obtained by both methods agree with those obtained by amino acid analysis of the a! subunit' (I 5).
Radioactivity Determinations--W and 3H activities were determined using a Nuclear-Chicago model 3000 nmbicnt temperature liquid scintillation counter operated at the balance point for each isotope.
Aqueous samples were counted with Bray's solution (30) from which 1,4-bis[2-(5~phenyloxa~zolyl)]benzene (POPOP) and ethylene glycol were omitted (31). Enzyme Assays-Activity of the a subunit, in stimulat,ing the conversion of indole and scrine to tryptophan by the f12 subunit was assayed as described by Smith and Yanofsky (32). The conversion of InGP to indole and triose phosphate by the (Y subunit was assayed in a reaction mixture containing 20 mar potassium phosphate buffer (pH 7.0), and 0.5 nnw InGP.
When the InGI' breakdown reaction was to be measured in the pi'csence of excess p2 subunit, the reaction mixture was made 30 mxf in hydroxylamine and 180 m&r in NaC'l (3). Indole formation was monitored either by the appearance of t,oluene-extractable radioactivity when 14C-InGP was used as a substrate, or by colorimctric assay of indole as described by Srnit'h and Yanofsky (32).
Treat?,lent of CI Subunit with BJIR, iVl~?lJ, and DTNB---Uefore treatment with these reagents, (Y subunit preparations were dialyzed exhaustively against 20 mM potassium phosphate buffer (pH 7.0). The cx subunit, 0.5 mg per ml, was treated with 0.5 mm NEM or BME in 20 mM potassium phosphate buffer (pI1 7.0) at 21-23" for 1 t,o 12 hours. The reaction was terminated by addition of mercaptoethanol in an-fold molar excess with respect to BME or NEM, and the sample exhaustively dialyzed against 20 mM pot,assiurn phosphate buffer (pH 7.0). After protein and radioactivity measurements, the moles of reagent bound per rnole of o( subunit were calculnt,ed. The DTNB titration met.hod originally described by Ellman (33) was used under conditions of higher reagent concentration and neutral pH similar to those recommended by Janatova,Fuller,and Hunter (34). Titrations were performed on 10 to 30 nmoles of o( subunit in 10 to 20 mM potassium phosphate buffer (pH 7.0). The reaction was initiated by the addition of 0.1 ml of a 10 mM DTNB solution made up in 0.1 M potassium phosphate buffer (pH 7.0). When DTNB titrations were performed in 4 to 5 M urea, the urea was added last. The final volume was 1.0 ml, and the optical density increase at 412 rnp was used in the calculation of the free thiol content of the protein as described by Ellman (33).
Enzymatic Digestions-TPCK-trypsin and chymotrypsin (Worthington) were used to make tryptic and tryptic-chymotryptic digests of the cr subunit as described by Helinski and Yanofsky (35). In tryptic-chymotryptic digestions, the protein was digested with TPCK-trypsin alone for 90 min, and then for an additional 90 min with both enzymes. Digests were freed of the 2 M urea-O.1 hf ammonium carbonate digestion buffer by gel filtration on columns of Sephadex LH-20 or G-25, and lyophilized.
Peptide Pur$cation-Tryptic digests of NEM-or BMEtreated LY subunit were chromatographed on columns, 133 x 0.9 cm, of Dowex 5OW-X2, 200 to 400 mesh, at 37". The peptides were eluted from the column by a convex gradient from 0.2 M pyridine, 5.0 M acetic acid, pH 3.1 (333 ml) to 2.0 M pyridine, 2.5 M acetic acid, pH 5.0 (666 ml), as described by Schroeder (36). Tryptic-chymotryptic digests were chromatographed on columns, 133 x 0.9 cm, of Dowex AG l-X2 at 37", with the procedures for resin preparation described by Birshstein, Hussain, and Cebra (37). l'eptides were eluted from the resin either with the complex pyridine acetate gradient described by Birshstein et al. (37) or by stepwise elution with 180 ml of 3% pyridinc, followed by 180 ml of 0.1 M pyridine, 0.05 M acetic acid (pH 5.5), and, finally, a linear gradient between 180 ml of the latter buffer and 180 ml of 2.0 M pyridine, 2.0 M acetic acid (pH 5.0). Radioactive peptide peaks were concentrated and pyridine acetate removed by flash evaporation.
Final purification of radioactive peptides was achieved by high voltage paper electrophoresis at pH 3.6 as described by Helinski and Yanofsky (35). l'eptides were eluted from the paper with 10% acetic acid.
Peptide Amino Acid Compositions-All samples analyzed were homogeneous as judged by analytical scale paper electrophoresis. Samples were hydrolyzed for 22 to 24 hours in a vacuum at 110" in 5.7 N HCl, and amino acid analyses performed with the Beckman model 120 B amino acid analyzers of the Department of Biology service facility. Peptide amino acid composition were calculated as follows. The amount of each amino acid found was first divided by the calculated input of radioactive 3H-BME or r4C-NEM peptide hydrolysate.
The actual amount of peptide hydrolysate analyzed on the column was then calculated by averaging the recoveries of those stable amino acids which clearly represented I-, 2-, or a-fold multiples of a value close to, and usually 10 to 3097, less than, the calculated amount of sample loaded onto the column, thus correcting for losses during transfer operations. As only limited amounts of material were available for analysis of most of the purified cross-linked peptides isolated, only 22-to 24.hour hydrolyses were performed, and the serine and threonine values were not corrected for losses caused by  E/E, is plotted on a logarithmic scale, and has been calculated assuming the stoichiometry of the BME reaction with 01 subunit is 1: 1.

oxidation.
Cysteine that had reacted with substituted maleimide reagents was determined as cysteinyl succinic acid (Xcysteinosuccinic acid). It has been observed that the yield of this derivative after 22-to 24.hour hydrolyses is approximately 50% (38, 39).

Properties of BME-treated 01 Subunit
The choice of conditions for reaction of the (Y subunit with BME was based on several well recognized facts. Neutral pH and low reagent concentrations favor the specific reaction of substituted maleimide reagents with the cysteinyl sulfhydryl groups of proteins, whereas reaction with the side chains of lysyl and histidyl residues, and with NHz-terminal amino groups, occurs under conditions of higher reagent concentration and more alkaline pH (39-41).
Furthermore, low concentrations of protein and of bifunctional reagent should favor bifunctional reactions leading to the formation of intramolecular, rather than intermolecular cross-links (42). When a! subunit is treated with BME as described under "Materials and Methods" for 6 hours, a maximum of 1 molecule of BME reacts per molecule of enzyme (Fig. l), and no further BME reacts with enzyme when treatment is continued for an additional 6 hours. Reagent is present in approximately 30.fold excess with respect to a! subunit under these conditions. The kinetics of the labeling process is pseudo first order, with a halftime of approximately 60 min (Fig. 1, inset). A 1: 1 ratio of moles of bifunctional reagent bound per mole of protein is compatible with the formation of a single intra-  (42), yet formation of doubly cross-linked dimers, or simple monofunctional reaction with each protein molecule could also yield such ratios.
A heterogeneous population of reacted and unreacted a! subunits might also fortuitously yield a net BME to OL subunit ratio of 1.
Sedimentation velocity studies of BME-treated Q! subunit eliminated the possibility that dimers or higher order aggregates were present.
Analytical ultracentrifugation with the Schlieren optics of the Spinco model E instrument, and analytical banding through 5 to 20% sucrose gradients showed that BME-treated cy subunit sedimented with a single symmetric refractive index gradient boundary and as a single peak of recovered 3H-BME radioactivity, respectively (Fig. 2). The calculated S values are identical with that of the native a! subunit monomer (4). BME-treated Q( subunit preparations were electrophoretically homogeneous and indistinguishable from the native a! subunit with the polyacrylamide disc gel electrophoresis system described by Malkinson and Hardman (15), which they found capable of separating enzymatically active, singly labeled species of OL subunit from inactive, multiply labeled species found after the reaction of the protein with NEM at pH 8.3.
To the limit of the resolving power of these hydrodynamic and electrophoretic criteria, the BME-treated Q! subunit preparations are free of intermolecularly cross-linked multimers, and are not detectably heterogeneous.
In the absence of dimer formation, the recovery of 2 moles of modified cysteine per mole of BME-treated Q! subunit would be strong evidence for an intramolecular cross-link between 2 cysteinyl residues. Total amino acid analyses of BME-treated Q! subunit after 24 and 48 hours of acid hydrolysis at 110" yielded 1.1 moles of cysteinyl succinic acid per mole of cy subunit. Since it has been reported that cysteinyl succinic acid is liberated in approximately 50% yield after 24 hours of acid hydrolysis of the reaction product of NEM and cysteine (38, 39), these data suggest that 2 cysteinyl thiols on each protein molecule had reacted with BME.
The difficulties associated with accurately determining small amounts of this derivative in the presence of large amounts of other amino acids suggest the more conservative interpretation that at least 1, and perhaps as many as 2 of the OL subunit cysteinyl residues have reacted with the reagent.
Similarly  In view of the possibility that BME might have reacted only monofunctionally with the (I( subunit, an experiment was designed to test for the presence of unreacted maleimide rings bound covalently to the a! subunit.
This was done by determining the extent to which free 14C-cysteine could be bound covalently to the enzyme in the presence of urea. The specific radioactivity of the 14C-cysteine added was sufficient to permit detection of 0.1 mole of cysteine bound per mole of LY subunit.
Results of such an experiment are shown in Fig. 3. During the period from 40 min to 6 hours after the initiation of BME treatment, no more than 0.1 to 0.2 mole of cysteine bound per mole of ac subunit could be detected.
However, the interpretation of these data is complicated by the fact that o( subunit which has never been exposed to BME (zero time point, Fig. 3) binds approximately 1 mole of cysteine per mole of a! subunit under these conditions. The mechanism of this cysteine binding is most probably either the oxidation of the free cysteine thiol to a mixed disulfide with one or more of the three a! subunit cysteinyl thiols, or a simple noncovalent binding of free cysteine to the enzyme.
In view of the exposure of the protein to urea during the cysteine treatment and the exhaustive dialysis afterwards, the first interpretation seems more likely.

Localization of Sites of Reaction
Sites of Reaction of Monojunctional Analogue, NEM-An a subunit preparation was treated with NEM for 12 hours under the same conditions used for BME treatment.
As shown in Fig. 4, treatment with the monofunctional reagent under these conditions proceeds at a rate similar to that of the BME cy subunit reaction.
However, if the time period is increased beyond 6 hours, more than a net of one NEM may react per mole of protein.
Accordingly a preparation labeled to the extent of 1.5 moles of NEM per mole of protein was chosen in hopes of gleaning the maximum amount of information about possible sites of reaction by the bifunctional reagent from monofunctional analogue studies.
This material was digested with TPCK-trypsin and the digest desalted with Sephadex LH-20 with 70% recovery.
The peak fractions containing 90% of the readioactivity recovered from the Sephadex LH-20 column were lyophilized and then chromatographed on Dowex 50 with 64% recovery. Two sharp well separated peaks, termed I and II, containing 49.5 and 50%, respectively, of the recovered radioactive NEM peptides, were obtained. The cysteine binding seen in the experiment described in High voltage paper electrophoresis at pH 3.6 showed that Fig. 3 may then represent residual oxidative or noncovalent each of these peaks was heterogeneous: each contained four well binding of cysteine, as well as the trapping of a! subunits that separated ninhydrin-positive bands, of which two were in each have reacted only monofunctionally with BME. The BME-case radioactive. The radioactive peptides from the first treated (Y subunit preparations must then contain at least 80 to Dowex peak pool (I-A, I-B) and the second Dowex peak pool 90% intramolecularly cross-linked monomer. It is also clear (II-A, II-B) were purified by preparative scale high voltage that the reaction of the second maleimide ring with the protein paper electrophoresis. On the basis of the amino acid composimust occur very rapidly after the initial reaction of the protein tions shown in Tables II and III, Peptides I-A and I-B were molecule with BME.
assigned to the region of the protein sequence surrounding  cysteine 80, and II-A and II-B to the region surrounding cysteine 117. No radioactive peptides could be found representing the reaction of NEM with cysteine 153. Attempts to recover additional radioactivity from the column by exhaustive washing were unsuccessful, and it is believed that the 20 to 30% losses during the purification steps represent nonspecific losses caused by the transfer and concentration steps involved. BME-Cross-linked Tryptic Peptides-A TPCK-trypsin digest of 3H-BME-treated protein was desalted with Sephadex G-25 with 73% recovery.
More than 50% of the radioactivity was eluted at the void volume of the column, suggesting that the cross-linked peptide fragments had molecular weights of at least 5000, although aggregation might also have occurred. This major radioactive fraction was lyophilized and chromatographed on Dowex 50 with 66% recovery.
The radioactive peptides were eluted in the early part of the gradient, emerging in four peaks: Peak A, a sharp peak near the void volume of the column containing less than 10% of the recovered radioactivity; Peak B, a broad zone containing some 169ib of the recovered radioactivity; and Peaks C, E, and F, three closely associated peaks each containing approximately 25yo of the recovered counts.
All radioactive fractions were then tested for heterogeneity with analytical scale high voltage paper electrophoresis, and, if needed, subjected to preparative scale purification by the same method.
The amino acid composition of the Peak A material was compatible with monofunctional BME-labeling of lysine 35, resulting in the production of a large tryptic peptide containing residues 16 to 69 of the a! subunit. Zone B and Peak F, even after purification attempts, gave amino acid analyses in which the ratios of most of the amino acids to the calculated input of BME peptide were extremely high, indicating that only very limited tryptic digestion had occurred. No specific peptide regions of the protein could be recognized on the basis of amino acid analyses of purified material from this part of the Dowex 50 chromatogram, and the quantities recovered were not sufficient to permit further digestion attempts with other proteolytic enzymes. The recovery of such large pieces of poorly digested reagent-treated protein was not observed with the NEM-treated QI subunit tryptic digests. The large fragments may represent interference with the normal course of tryptic digestion of the LY subunit resulting from the introduction of the intramolecular cross-link. The peptides from Peaks C and E of the Dowex 50 chromatogram are large but nonetheless recognizable.
A rigorous comparison of the composition of these peptides with all known tryptic peptide compositions indicate that these radioactive peptides result from a BME cross-linkage between cysteines 80 and 117 (Table IV).
The results obtained after tryptic and chymotryptic digestion bear out this conclusion. BME-cross-linked Tryptic-Chymotryptic Peptides--In three separate experiments BME-treated (Y subunit preparations were digested with trypsin and chymotrypsin in attempts to decrease the losses and size heterogeneity associated with the BME subunit tryptic peptides.
Over-all yields on desalting by Sephadex G-25 chromatography were approximately 90 %, and nearly all radioactivity was eluted as a single peak at the void volume of the column.
This material was pooled, lyophilized, and chromatographed on Dowex 1. Over-all recovery from the   Dowex 1 columns was 90% or better, with 95% of the recovered radioactive peptide material emerging as three incompletely separated peaks in the early part of the gradient. Radioactive peptide material was purified further by high voltage paper electrophoresis.
In some experiments, material from the Dowex 1 peak pools was subjected to performic acid oxidation by the method of Him (43) before paper electrophoresis.
No additional radioactive or ninhydrin-positive spots could be found when the electropherograms of performic acid-oxidized samples were compared with those of untreated samples, nor were the amino acid compositions or electrophoretic mobilities of the isolated radioactive peptides detectably altered.
Performic acid oxidation of the cysteinyl sulfur involved in a BME cross-link (by addition of the thiol group across the maleimide double bond) would be expected to produce a sulfone or sulfoxide, but not to result in cleavage of the thioether. Hence a BME cross-link would be stable to performic acid oxidation. basis of these compositions, the cross-link represented by these peptides is between cysteines 80 and 117. No peptides were found which give amino acid analyses compatible with the involvement of cysteine 153 in a BME cross-link. Moreover, the BME tryptic-chymotryptic peptides all appear to be smaller fragments of the large tryptic BME ac subunit peptides presented in Table IV, which in turn approximate the sum of the cysteines 80 and 117 NEM-reacted peptides as presented in Tables II  and III. Catalytic and Immunochemical Properties of NEM-and BME-treated OL Subunits The relative specific activity of the CY subunit treated with either BME or NEM falls to approximately 30% of the level found in untreated control preparations in all of the reactions tested, including not only reactions carried out by the unassisted CY subunit but also reactions carried out by the CY& complex.
The time course of BME inactivation of the CY subunit's enzymatic activity in the conversion of InGP to indole and triose phosphate in the presence of excess /32 subunit is shown in Fig. 5 appears to be pseudo first order with respect to the concentration of active OL subunit. As the fractionally inactivated NE% and BME-treated enzyme preparations were apparently homogeneous with respect to molecular weight and electrophoretic properties, the nature of the fractional remaining enzymatic activity was examined further with the method of Creighton and Yanofsky (3) for studying the association of the a: and & subunits by sucrose density gradient ultracentrifugation.
On sucrose density gradient centrifugation, in the presence of excess /$ subunits, both NEM-and BME-treated ac subunits showed a distinct heterogeneity.
Only 30 to 40% of the treated (Y subunits applied to the gradient were able to combine with the & subunit (Fig. 6). The remaining 60 to 70% sedimented in the same position as untreated a-subunit monomers in the absence of the p2 subunit.
The failure of a fraction of the treated material to form a functional tryptophan synthetase complex with the pZ subunit suggested that immunochemical techniques might be useful in the detection of structural heterogeneity in the BME-and NEM-treated o( subunit preparations. If the tertiary structure of the inactive two-thirds of the treated (II subunit population were grossly altered, it would be expected that immunochemical reactivity with antibody prepared against normal (Y subunit would also be lost.
If only the enzymatically active molecules in the treated population were immunochemically similar to the normal cy subunit, then only 1 unit of antibody to normal 01 subunit would be required to neutralize a unit of enzymatic activity in the treated preparation.2 Fig. 7A shows that this is not the case, and that at least 2 antibody units are required to neutralize 1 enzyme unit in the treated material. This indicates the presence of a significant fraction of 01 subunits which retain immunochemitally reactive tertiary structure, although they are catalytically inert.
Precipitation of normal and treated (Y subunits in the presence of excess antibody was also studied. As shown in Fig. 7B, the quantitative precipitin reaction curves for all three types of a: subunit, normal, NEM-treated, and BME-treated are superimposable. This is further evidence of a high degree of structural similarity between the normal and modified protein species. Aliquots of these incubation mixtures were assayed for a subunit activity in stimulating the indole to tryptophan.reaction (excess pz subunit).
Symbols: 0, normal o( subunit; A, BME-treated 01 subunit. The solid line has a slope of 1 (1 antibody unit required to neutralize 1 tryptophan synthetase unit), and the dashed line a slope of 0.5 (2 antibody units required to neutralize 1 tryptophan synthetase unit).
B, quantitative precipitation of treated and normal (Y subunits by excess antinormal L\! subunit antibody.
Increasing amounts of normal and modified 01 subunit were incubated with a constant excess of antibody for 3 hours at room temperature, and for 48 hours at 4'. The reaction tubes were then centrifuged for 10 min at 10,000 X g in a refrigerated centrifuge.
The precipitates were resuspended once in 20 mu potassium phosphate buffer (pH 7.0) containing 0.15 M NaCl. recentrifuzed. and finallv resusnended in s final volume of 1.0 ml of the same 'buffer. The amount of precipitate formed was quantitatively determined by measurement of the turbidity at 550 rnp (45). Symbols: 0, normal (Y subunit; A, BME-treated LY subunit; n , NEM-treated a! subunit.

DISCUSSION
Clearly the best evidence for production of a unique intramolecular cross-link by a bifunctional reagent would be the isolation of a single cross-linked peptide in quantitative yield. In this case, however, the conclusion that only cysteines 80 and 117 of the tryptophan synthetase (Y subunit react with BME to yield an intramolecular cross-link has been based on two lines of evidence. First, studies with an appropriate monofunctional maljimide reagent, NEM, indicated that under the conditions chosen only cysteines 80 and 117 were available for reaction with this type of reagent.
The fact that enzyme molecules could react with NEM to the extent of a net of 1.5 moles of NEM per mole of a! subunit indicated that both cysteines 80 and 117 might react on the same molecule.
Second, it was possible to isolate from TPCK-trypsin digests of aH-BME-labeled o( subunit, in addition to large peptides or protein fragments which indicated poor proteolytic digestion of the cross-linked a! subunit, peptides clearly recognizable as cysteine 80 to cysteine 117 cross-linked peptides. Although these findings indicated that the cysteine 80 to cysteine 117 cross-link was indeed formed, the low yields of the recognizable peptides found suggested that other cross-links might also be present.
In addition, tryptic-chymotryptic digests of 3H-BME-treated a! subunit gave high yields of two types of smaller peptides which could be assigned to the cysteine 80 to cysteine 117 cross-link.
These peptides moreover appeared to be smaller fragments of the largest 3H-BME-treated (Y subunit tryptic peptides that could be recognized.
The observed monofunctional reaction of BME with lysyl residue 35 in the 3H-BME-treated a! subunit preparation used in the tryptic digestion experiments may be an anomaly, as reaction of this residue with either NEM or BME could not be detected in any of the other experiments.
On the basis of the electrophoretic, hydrodynamic, and immunochemical properties of the BME-treated a! subunit preparation, and the finding that all of the intramolecular cross-links are apparently between cysteines 80 and 117, the heterogeneity detected by enzymatic methods and by the sucrose density gradient studies of interaction of the 01 and & subunits remains puzzling.
The BME-and NEM-treated (Y subunit preparations are very similar to the normal tryptophan synthetase c11 subunits in many aspects of their structure.
The differences observed are the consequences of the reaction of these bifunctional and monofunctional sulfhydryl reagents with cysteine residues 80 and 117.
The generation of functional heterogeneity in cysteine 80 to cysteine 117 cross-linked cx subunit might occur by one of the following two mechanisms, or possibly by both acting in concert.
The addition of the thiol group of L-cysteine to the olefinic double bond of NEM has been shown to produce a new asymmetric center (39). Two diastereoisomeric forms of the product (X-l-ethyl-Z, 5 dioxopyrrolidin-3-yl)-L-cysteine are produced. Extrapolation of these findings to the case of a symmetrical bifunctional maleimide reagent with free rotation between the maleimide rings leads to the conclusion that the addition of two protein cysteinyl thiols across the maleimide olefinic bonds must generate two new centers of asymmetry in the cross-link itself. Hence as many as four conformationally distinct crosslinked protein species might be formed, and the conformational heterogeneity thus introduced at or near the active site might be responsible for the functional heterogeneity observed.
A second mechanism which may possibly act together with the first involves the order of reaction of the two thiols with the bifunctional reagent. Although an experiment was performed with the purpose of trapping BME treated intermediates in which only one of the BME maleimide rings had reacted with an a! subunit thiol, no intermediates in the cross-linking process could be detected.
The reaction of the second maleimide ring must then follow quite rapidly after the initial reaction.
It is not known whether the initial reaction of BME is with cysteine 80 or 117, or whether the initial reaction of BME may be at either site. It is possible that the site of the initial reaction may be a factor in generating the functional heterogeneity observed. Initial reaction at one or the other of the two sites might affect local tertiary structure of the protein or influence the stereochemical course of the reaction of the second maleimide with its target sulfhydryl.
The order of addition of the two thiols to the BME maleimide rings might then not only affect the conformation of the protein in the immediate area of the cysteine