Purification and Characterization of 0-Acetylserine Sulfhydrylase Isoenzymes from Datura innoxia*

Three isoenzyme forms (designated A, B, and C) of 0-acetylserine sulfhydrylase were purified from Datura innozia suspension cultures. Isoenzyme A is the most abundant form, comprising 45430% of the total activity. Isoenzymes C and B comprise 3540% and 10-20% of the activity, respectively. The specific activities of the purified isoenzymes are similar (870-893 pmol of cysteine/ midmg of protein). Molecular masses for isoenzymes A, B, and C, estimated by analytical size exclusion high performance liquid chromatography, are 63,86, and 63 kDa, respectively. Isoenzymes A and B are homodimers; isoenzyme C is a heterodimer. Spectral analysis indi- cates that these isoenzymes possess a pyridoxal 5’-phos-phate cofactor that binds the 0-acetylserine substrate. Binding is reversible by addition of the sulfide substrate. The 0-acetylserine sulfhydrylase isoenzymes are active over a broad temperature range, with maximum activity between 42 and 58 “C. They are active only be- tween pH 7 and 8, with optimal activity at pH 7.6. Kinetic analysis indicates these enzymes are allosterically regu- lated and exhibit positive cooperativity with respect to

caused by a number of stress and injury conditions, including presence of toxic metals, heat shock, and radiation (for reviews, see Larsson et al. (1983); Meister and Anderson (1983); Reed and Fariss (1984); Deneke and Fanburg (1989); and Fahey and Sundquist (1991)). Synthesis of glutathione and its precursors may be induced by conditions causing physiological stress.
The presence of 0-acetylserine sulfhydrylase activity in plants was documented during the 1960s and 1970s (Smith and Thompson, 1971Masada et al., 1975;Ascaiio and Nicholas, 1977;Bertagnolli and Wedding, 1977). However, much of this earlier work was conducted with crude plant extracts, or the purity of the enzyme preparations was not demonstrated. Improvements in separation chemistry have allowed us to obtain highly purified enzyme preparations from plants, and to initiate a more critical evaluation of this enzyme. Plants contain multiple isoenzyme forms of 0-acetylserine sulfhydrylase that are located in the organelles and the cytosol (Bertagnolli and Wedding, 1977;Ikegami et al., 1987;Nakamura and Tamura, 1989;Lunn et al., 1990;Droux et al., 1992;Rolland et al., 1992). The 0-acetylserine isoenzymes from Datura innoxia probably perform different metabolic functions in the plant cell and respond differently to different stimuli or requirements for cellular metabolites.
We have purified three 0-acetylserine sulfhydrylase isoenzymes from D. innoxia and are investigating the roles that these three isoenzymes play in plant cell metabolism, particularly with respect to sulfur assimilation and stress-induced use of cysteine in biosynthesis of glutathione and its metal-binding derivatives (Delhaize et al., 1989;Jackson et al., 1992;Jackson and Kuske, 1993). We present here a comparison of their physical and in vitro catalytic properties.

Reagents
Except where noted, reagents for enzyme assays, buffers, protein purification, and silver staining were the highest punty obtainable and were purchased from Sigma. Chromatography resins were purchased from Pharmacia LKI3 Biotechnology Inc. Coomassie dye-binding reagent and protein standards were obtained from Bio-Rad, and the BCA (bicinchoninic acid) assay kit was obtained from Pierce Chemical Co.

D. innoxia Cell Suspension Cultures
Cadmium-tolerant (Cd 160R and Cd 3009 and -sensitive (wild-type WDI and a cadmium-sensitive revertant of Cd 300R, Cd 3OOs) D. innoxia cell suspension cultures were grown in the dark at 30-33 "C in a modified 1B5 medium as described by Jackson et al. (1984). Plant cells for protein purification were grown as 100 ml batch shake cultures in 500-ml De Long flasks at 30-33 "C. Cultures were maintained by dilution of one part cell culture to three parts fresh medium every 3 days. This induced a burst of cell growth. The cadmium-tolerant cells are able to grow continuously in medium containing 250 mM CdCl,, while the sensitive cell lines die within 48 h following exposure to this CdCl, concentration (Delhaize et al., 1989). 0-Acetylserine Sulfhydrylase Purification Soluble Protein Extraction-Due to equipment and space constraints, we purified the 0-acetylserine sulfhydrylase proteins in multiple batch runs, each beginning with 250-800 g wet weight of cells from 1-1.5 liters of original growth medium. Cadmium-tolerant D. innoxia (Cd 3009 were harvested 3 days after transfer to fresh medium. Cells were collected by centrifugation at 1000 x g for 1 min using a tabletop swinging bucket centrifuge. Growth medium was decanted from the pelleted cells and replaced with an equal volume of extraction buffer (30 m~ Tris, pH 7.8, 10 m~ 6-mercaptoethanol). Cells were thoroughly resuspended in this buffer and pelleted as above. The buffer was aspirated from the pelleted cells, and the wet weight of cells was recorded. Pelleted and rinsed cells were quickly frozen in liquid nitrogen and ground to a fine powder. The powder was resuspended in 50 "C extraction buffer and cooled in an ice water bath. All subsequent procedures were carried out at 0-4 "C. The brei was centrifuged at 14,000 x g for 20 min, and the supernatant was collected. Batch preparations contained between 2.5 and 5 g of soluble protein. Preliminary experiments indicated that the 0-acetylserine sulfhydrylase activity was associated with the soluble protein fraction.
slowly added to the supernatant with stirring to 20% saturation and the Ammonium Sulfate Precipitation-Granular ammonium sulfate was precipitate was collected by centrifugation at 21,000 x g for 20 min. The pellet was discarded and the supernatant was slowly brought to 80% ammonium sulfate saturation. The precipitated protein was pelleted by centrifugation at 21,000 x g for 20 min and gently resuspended in 20-60 ml of 60 m~ Tris, pH 7.8, 10 m M P-mercaptoethanol, 10% glycerol.
Gel Filtration Chromatography-Ten-to 12-ml aliquots of the protein sample were applied to Sephacryl200-HR columns (100 x 2.5 cm) equilibrated in 60 m~ Tris, pH 7.8, 10 m~ P-mercaptoethanol, 10% glycerol. 0-Acetylserine sulfhydrylase-active fractions eluted as a broad single peak following the void volume protein peak that contained most of the protein. Fractions with high specific activity were pooled and concentrated by filtration using a 30-kDa exclusion filter (Amicon filtration unit).
Anion Exchange Chromatography-The concentrated protein sample was diluted to 10 m~ Tris and the pH adjusted to 8.1. It was then loaded onto a DEAE CL-GB column (30 x 5 cm) equilibrated in 10 m~ Tris, pH 8.1. The column was rinsed with 100 ml of the above buffer, and protein was eluted in a 200-ml linear gradient of 50-250 m~ NaCl in 10 m~ Tris, pH 8.1.0-Acetylserine sulfhydrylase-active fractions eluted as two peaks, a broad major peak that began eluting at 120 m~ NaCl and a minor peak eluting at approximately 230 m~ NaC1. These peak fractions were designated 0-acetylserine sulfhydrylase isoenzymes A and B, respectively. Fractions containing the A and B peaks were pooled (separately) and concentrated by centrifugal filtration at 4000 x g through 30-kDa exclusion filters (Centriprep-30 concentrators, Amicon).
Size Exclusion HPLC'"Isoenzymes A and B were each further purified by chromatography through a preparative BIO-SIL SEC-250 FPLC column (600 by 21.5 mm), equilibrated in 70 m~ Tris, pH 7.8, using a Waters dual pump, programmable HPLC system. Two-ml fractions of each isoenzyme with high specific activity were pooled, and buffer was exchanged into 10 m~ Tris, pH 8.1, using Centriprep-30 concentrators as described above.
Anion Exchange HPLC-Concentrated aliquots of A and B were applied to a Bio-Gel SEC DEAE-5-PW column (75 x 7.5 mm) equilibrated in 10 m~ Tris, pH 8.1. Protein was eluted in a 60-ml linear gradient of 0-300 m M NaCl in 10 m~ Tris, pH 8.1, at a flow rate of 1 ml/min. When isoenzyme A was applied to the column, 0-acetylserine sulfhydrylase activity eluted as two sharp peaks, the first (isoenzyme A) at 120 m~ NaCl and the second (isoenzyme C) at 140 m~ NaCl. Isoenzyme B eluted as a single sharp peak at 245 m~ NaCl and at this point was apparently pure.
Nondenaturing Acrylamide Gel Electrophoresis-Isoenzymes A and C were further purified by electrophoresis through 7.5% nondenaturing acrylamide gels. Adjacent test slices of the gel were either stained with silver to locate the separated proteins, or assayed for 0-acetylserine sulfhydrylase activity following electroelution of the protein from the test gel slice using a Centrilutor microelectroeluter and Centricon-30 microconcentrators (Amicon). The region of the gel containing the 0acetylserine sulfhydrylase protein was excised, and the protein was electroeluted from the gel using an Elutrap device (Schleicher & Schd l ) . Purified isoenzyme A, B, and C proteins were stored in 1&30 m~ Tris, pH 8.1, 10% glycerol at -70 "C. The above purification procedure was repeated several times to obtain sufficient protein for investigation of the physical and kinetic characteristics of the three enzymes.
Determination of Protein and Enzyme Purity Protein concentration was determined by the method of Bradford (1976) (Bio-Rad protein assay dye) and a modified Lowry assay (Lowry et al., 1951) (BCA assay, Pierce Chemical Co.) using bovine serum albumin as the standard. Purity of the three 0-acetylserine sulfhydrylase isoenzymes was determined by one-and two-dimensional SDS-polyacrylamide gel electrophoresis (OFarrel, 1975;Hames, 1981) and electrophoresis in a series of nondenaturing acrylamide gels (7.515% acrylamide) (Hames, 1981). Gels were stained with silver using the procedure described by Blum et al. (1987).
Presence of 0-Acetylserine Sulfhydrylase Isoenzymes in Cadmium-sensitive a n d -tolerant D. innoxia Cell Cultures D. innoxia cells from two cadmium-tolerant cultures (Cd 300R, Cd 1609 growing in 250,160, or 0 m~ CdCl,, the wild type cell line (WDI), and a cadmium-hypersensitive cell line (Cd 3OOs) were harvested and soluble proteins obtained as described above. Twenty mg of protein samples were applied to a Bio-Gel SEC DEAE-5-PW column (75 x 7.5 mm) in 10 m~ Tris, pH 8.1, and eluted in a 60-ml linear gradient of &300 m~ NaCl in 10 m~ Tris, pH 8.1. One-ml fractions were assayed for 0-acetylserine sulfhydrylase activity.
Native Molecular Mass and Subunit Structure of the 0-Acetylserine Sulfhydrylase Isoenzymes Native molecular masses of the three 0-acetylserine sulfhydrylase isoenzymes were estimated using gel filtration HPLC on an analytical Bio-Si1 SEC-250 column (300 x 7.8 m m ) equilibrated in 70 m~ Tris, pH 7.8. Purified isoenzymes were applied to the column separately and as mixtures of A plus B or A plus C, along with the following protein standards: thyroglobulin, 670 kDa; IgG, 158 kDa; ovalbumin, 44 kDa; myoglobin, 17 kDa; and vitamin B-12, 1.35 kDa. Flow rate was 1 ml/ min. Protein elution was monitored by absorbance at 280 nm, and 2 0 0 4 fractions were collected and tested for 0-acetylserine sulfhydrylase activity as described above. Molecular mass (m) was calculated from the linear relationship between In m and relative elution volume (VJ V,,). Determinations were repeated twice for each isoenzyme. Subunit structures were determined by SDS-polyacrylamide gel electrophoresis of purified proteins under reducing conditions (Hames, 1981). Molecular mass standards used in SDS-polyacrylamide gels are listed in Fig. 2.
Amino Acid Analysis of 0-Acetylserine Sulfhydrylase Proteins Amino acid compositions of the three 0-acetylserine sulfhydrylase isoenzymes were determined using a Pico-Tag amino acid analysis system (Bidlingmeyer et al., 1984;Cohen et al., 1989). Purified proteins were hydrolyzed under vacuum in constantly boiling HCl at 112 "C for 45 h and derivatized with phenylisothiocyanate. PTC-derivatized amino acids were separated by HPLC on a calibrated Waters (3-18 Pico-Tag hydrolysate column (15 cm x 3 mm) at 41 "C. The presence of tryptophan was not determined. Cysteine was present in all three proteins but was not derivatized prior to hydrolysis and therefore was not quantified.

WIVisible Spectral Studies of 0-Acetylserine Sulfhydrylase Isoenzymes
Absorbance spectra were measured in a Perkin-Elmer Lambda 3B W M S dual beam spectrophotometer, recording at 120 n d m i n and 10 ndcm, using a 1-cm path length. The 0-acetylserine sulfhydrylase proteins were analyzed in either 10 m~ Tris, pH 8.0 (Fig. 4) or 50 m~ Tris, pH 7.6 ( Fig. 5), and the spectrophotometer was blanked using the respective buffer solution. Pyridoxal 5"phosphate was quantified by comparison of absorbance at 412 nm to a standard curve of known concentrations in 10 m~ Tris, pH 8.0.
0-Acetylserine Sulfhydrylase Activity Assay 0-Acetylserine sulfhydrylase activity was assayed by measuring the production of L-cysteine. For initial steps of enzyme purification, 1-ml reaction mixtures contained 100 m~ Tris, 20 m~ 0-acetylserine, 1 m~ Na,S, at a reaction pH of 7.6. 0-Acetylserine and NazS were freshly prepared as 1 M and 100 m~ stocks, respectively, in 100 m~ Tris, pH 8.0, just prior to use. Reaction mixtures were preincubated at 34 "C for 2 min, and reactions were initiated by addition of 1-10 pl of protein sample. Each reaction was allowed to proceed at 34 "C for exactly 10 min. Reactions were stopped by addition of 200 pl of 1.5 M trichloroacetic acid. Precipitated protein was pelleted by centrifugation for 5 min in a microcentrifuge.
The amount of cysteine present in the supernatant was determined as described by Gaitonde (1967). One ml of the supernatant was transferred to a glass tube and mixed well with an equal volume of acid ninhydrin reagent (250 mg of ninhydrin in 10 ml of 3:2, v:v, glacial acetic acidconcentrated HCI). The mixture was heated in a boiling water bath for exactly 5 min and then cooled on ice. Two ml of cold 100% EtOH were added to each reaction and mixed well. The cysteine concentration was determined spectrophotometrically by absorbance at 546 nm. This assay is linear for L-cysteine at concentrations between 0.01 and 0.5 pmol. The 0-acetylserine sulfhydrylase activity assay was optimized with respect to pH, temperature, and concentration of each component and was linear with respect to the amount of enzyme assayed. One unit of activity is 1 pmol of cysteine/min/mg of protein. For purified and partially purified 0-acetylserine sulfhydrylase samples (after size exclusion FPLC), the NazS concentration in reaction mixtures was reduced to 200 p, after determining that this substrate is inhibitory to the purified enzyme at higher concentrations (see below).

Substrate Saturation Experiments
One-ml reaction mixtures contained 100 m~ Tris, different concentrations of 0-acetylserine and Na2S, and enzyme at a final reaction pH of 7.6. A 1 M stock of 0-acetylserine was prepared in 100 m~ Tris, pH 8.0, immediately before use and diluted as needed in the same buffer just before addition to the reaction mix. Na,S was prepared immediately before use as a 100 m~ stock in 100 IIIM Tris, pH 8.1, and diluted to the appropriate concentration immediately before addition to the reaction mix. The concentration of 0-acetylserine was varied between 0 and 40 m M at 200 PM Na,S. Na,S concentration was varied between 0 and 1200 p at 20 m~ 0-acetylserine. Twenty substrate concentrations were tested for each series, with data points concentrated at the lower substrate levels (Figs. 7 and 8). Reactions were performed as described above and initiated by the addition of 90 ng (in 1-3 pl volume) of purified enzyme sample. Each reaction was individually timed and allowed to proceed at 34 "C for exactly 10 min. Reactions were stopped and analyzed for L-cysteine as described above. Assays were linear with respect to cysteine production over 30 min. Substrate saturation experiments were repeated three to five times for each 0-acetylserine sulfhydrylase isoenzyme.

Effects of Temperature and pH on Enzyme Activity
Reaction mixtures for pH and temperature experiments contained 200 p NazS and 20 m~ 0-acetylserine in 100 m~ Tris buffer. For pH curves, the reaction mix pH was measured at 34 "C in parallel samples (minus enzyme). Reactions were initiated by addition of 1 p1 of enzyme in 10 m~ Tris, pH 7.6. The effect of different incubation temperatures on enzyme activity was examined by preincubating reaction mixes for 5 min at the desired temperature, adding enzyme, and continuing incubation at that temperature for 10 min. Reactions were stopped and analyzed for cysteine as described above. All pH and temperature experiments were repeated three times for each 0-acetylserine sulfhydrylase isoenzyme.

Effects of Cadmium and Other Metals on Activity
The presence of heavy metals induces biosynthesis of a specific class of metal-binding polypeptides in D. innoxia cell cultures (Robinson et al., 1988;Delhaize et al., 1989;Jackson et al., 1989;Rauser, 1990).
These metal-binding polypeptides are derived from glutathione, and their synthesis requires synthesis of glutathione and its amino acid precursors, cysteine, glycine, and glutamate. We explored the possibility that 0-acetylserine sulfhydrylase activity might be influenced by the presence of certain metal ions. The effects of cadmium, copper, zinc, iron, cobalt, and nickel on 0-acetylserine sulfhydrylase isoenzyme A were determined using the above standard assay. Metal concentrations of 0.01, 0.1, and 1 p had no effect on enzyme activity. All metals were inhibitory at 10 p~.

Analysis of in Vitro Kinetic Data
Data from replicate experiments on each isoenzyme were pooled and analyzed by non-linear least squares regression fitting to three models.
The Hill equation (Hill, 1925;Roberts, 1977;Ferscht, 1985) is where u is the velocity of the reaction (rate of product formation), V,,, is the maximum velocity, K is the rate constant, [SI is the substrate concentration, and n is that apparent kinetic order of the velocity with respect to S as S approaches zero (Hill number).
Parameter estimates for the three non-linear equations were derived using the Gauss-Newton method (Kennedy and Gentle, 1980) for a least-squares fit within the XLisp-Stat package (Tierney, 1990). Parameterizations of the equations were, as suggested by Ratkowsky (1983Ratkowsky ( , 1986Ratkowsky ( , 1990, to increase the chance of convergence. Analysis of O-acetylserine substrate saturation curves for the three isoenzymes included all data points. Analysis of sodium sulfide saturation curves was limited to sodium sulfide concentrations of 0-200 p for isoenzymes A and C and C150 p for isoenzyme B to avoid the inhibition portion of the curve (Fig. 8).

0-Acetylserine Sulfhydrylase Zsoenzymes in D. innoxia-The
same three isoenzymes were separated by DEAE-5PW HPLC from crude soluble proteins extracted from metal-tolerant and -sensitive D. innoxia cell cultures and from metal-tolerant cultures growing in the presence or absence of cadmium (Fig. l).
Isoenzyme A was the most predominant form of the enzyme in all of the D. innoxia cultures, comprising 4540% of the total activity. Isoenzyme C comprised 3540% and isozyme B comprised 10-20% of the total activity. The purified isoenzyme forms were identical to those detected in crude plant cell extracts, indicating that they are not artifacts of purification. Purification of Three 0-Acetylserine Sulfhydrylase Zsoenzymes (Table Z t P l a n t cell cultures used for 0-acetylserine sulfhydrylase purification were non-green, nonphotosynthetic cells, and thus did not possess the high concentrations of enzymes involved in photosynthesis in green tissues. In preliminary experiments, we found that 0-acetylserine sulfhydrylase activity was present in a broad range of ammonium sulfate concentrations. This is probably due to the presence of multiple isoforms. Although little purification was achieved at this step, ammonium sulfate served to concentrate and stablilize the enzyme. 0-Acetylserine sulfhydrylase isoenzymes co-eluted from the Sephacryl 200-HR column as a single broad peak eluting after the void peak that contained most of the protein. In repeated batch runs, no 0-acetylserine sulfhydrylase activity was detected in the void protein peak. Isoenzymes A and B were separated by anion exchange chromatography in a linear gradient of 50-250 mM NaCl in 10 mM Tris, pH 8.1. Isoenzyme A eluted with the bulk of the protein at about 120 n m NaCl, and B eluted at about 230 mM NaCl. Anion exchange HPLC on a Bio-Gel SEC DEAE-5-PW column purified isoenzyme B (Fig.   2B), and separated isoenzyme C from isoenzyme A. Nondenaturing acrylamide gel electrophoresis was used to separate isoenzymes A and C from a few remaining contaminant proteins (Fig. 2 , A and C ) . The three highly purified 0-acetylserine sulfhydrylase isoenzymes from D. innoxia had similar specific activities (for example, ranging from 870 to 893 units in the illustrated batch purification).
The final proteins were of very high purity as determined by electrophoresis through SDS-polyacrylamide gels (Fig. 2), twodimensional gels, and nondenaturing gels (data not shown).  Purified proteins were stable (i.e. no loss in specific activity) for 6 months when stored at -70 "C in 10-30 mM Tris, pH 8.1, containing 10% glycerol. They readily degraded when stored at 4 or -20 "C in buffered solutions, and with repeated freezingthawing. Isoenzyme B was less stable than isozymes A and C, possibly because it was stored at a less concentrated solution.
Addition of the pyridoxal 5"phosphate cofactor to the purification buffers or activity assay did not improve enzyme recovery or enzyme activity. At saturating 0-acetylserine levels, additional pyridoxal 5'-phosphate had no effect on enzyme activity. At subsaturating levels of 0-acetylserine, addition of 100 mM pyridoxal 5"phosphate reduced 0-acetylserine sulfhydrylase activity to 81% at 10 mM 0-acetylserine and to 50% at 0.5 m~ 0-acetylserine (data not shown).

Native Molecular Mass and Subunit Composition-
The native molecular mass for both isoenzyme A and C estimated by size exclusion HPLC was 63 kDa (Fig. 3). This value falls in the mid-range of native molecular masses reported for O-acetylserine sulfhydrylase proteins from other plants (Masada et al., 1975;Tamura et al., 1976;Bertagnolli and Wedding, 1977;Murakoshi et al., 1985;Ikegami et al., 1987Ikegami et al., , 1988 Boronat et al., 1984). In addition to these abundant forms of 0-acetylserine sulfhydrylase, we purified a less abundant form, isoenzyme B, that has similar specific activity to the other two isoforms, but has a molecular mass of 86 kDa (Fig. 3). This is the first report of an 0-acetylserine sulfhydrylase protein this large. The 0-acetylserine sulfhydrylase proteins that have been described from plants, bacteria, and an alga are homodimers.
Under reducing, denaturing electrophoresis conditions, the D. innoxia isoenzyme A ran as a single band a t 32 kDa, indicating that the protein is a homodimer (Fig. 2A ). Similarly, isoenzyme B ran as a single band at 43 kDa, indicating that it also is a homodimer (Fig. 2 B ) . Purified isoenzyme C consistently produced two bands of equal intensity. The larger species was 32 kDa in size, and the second species was about 800 Da smaller. To determine whether this second protein band was a contaminant protein, samples of purified isoenzyme C were subject to electrophoresis through a series of nondenaturing acrylamide gels with acrylamide concentrations of 5, 7.5, 10, 12, 15, and 17.5% and silver-stained. The native protein was visible as a single band in all gels. The single protein band present in nondenaturing gels consistently gave rise to two silver-stained bands on SDS-polyacrylmide gels, after electroelution of the native protein from the nondenaturing gel. These results indicate that isoenzyme C is a heterodimer, a subunit structure that is unique among the 0-acetylserine sulfhydrylase proteins that have been described.
Amino Acid Composition-The amino acid profiles of the three isoenzymes are similar. Histidine and serine contents varied between the isoenzymes, but there was little variation in the other amino acids (data not shown).
Spectral Properties of Purified 0-Acetylserine Sulfhydrylase Proteins-Absorbance spectra for the three D. innoxia isoenzymes over wavelengths 190-850 nm were very similar with identical absorption maxima (Fig. 4). The spectra of the purified enzymes exhibited two major peaks with absorbance maxima a t 278 nm, due to aromatic amino acids in the protein, and a t 412 nm. The absorbance at 412 nm is characteristic of the aldoxime form of pyridoxal 5"phosphate (Cook and Wedding, 1976; Kallen et al., 1985) and is due to a Schiff base formed between the pyridoxal 5'-phosphate cofactor and the €-amino group of a n enzyme lysine. Concentrated solutions of 0-acetylserine sulfhydrylase protein are yellow due to the presence of this cofactor. In general, bound pyridoxal 5'-phosphate is present in many transaminase enzymes (Wilson and Crawford, 1965;Kaplan and Flavin, 1966;Schnackerz et al., 1979), with characteristic absorbance between wavelengths of 400 and 430 nm (Zeffren and Hall (1973) and Cook et al. (1992), for examples). This cofactor has also been identified in O-acetylserine sulfhydrylase from bacteria and other plants Masada et al., 1975;Tamura et al., 1976;Murakoshi et al., 1985;Ikegami et al., 1987Ikegami et al., ,1988Drow et al., 1992). The molar ratio of pyridoxal 5'-phosphate (A412) to protein (AzTR) was 2:1, suggesting that there is 1 molecule of pyridoxal 5'- phosphate associated with each protein subunit. The O-acetylserine sulfhydrylase protein appeared to be saturated with tightly bound pyridoxal 5'-phosphate, since the activity of the pure enzyme was not increased by addition of free cofactor. The effects of the two substrates, 0-acetylserine and sulfide, on the absorbance spectrum of isoenzyme A were determined. The enzyme (28 pg/ml in 50 mM Tris, pH 7.6) was first titrated with 0-acetylserine at concentrations of 0.2, 1.0, 2.0, 4.0, 10.0, and 20.0 p~. Addition of the 0-acetylserine substrate to the purified enzyme results in considerable change in the absorbance spectrum, especially with regard to the absorbance due to pyridoxal 5'-phosphate. In the presence of 0-acetylserine at pH 7.6, there is a concentration-dependent spectral shift from 412 to 460 nm, accompanied by an increase in absorbance as a broad shoulder around 330 nm (Fig. 5). The enzyme preparation was saturated at 4.0 p~ 0-acetylserine. A similar spectral shift was observed in the pyridoxal 5"phosphate containing enzyme, D-serine dehydratase, by Schnackerz et al. (1979). They demonstrated that this type of spectral shift was due to the formation of a Schiff base intermediate between a-aminoacrylate and the pyridoxal 5"phosphate cofactor. Formation of this a-aminoacrylate intermediate was reversible upon addition of the second substrate, sulfide. Similarly, the observed changes in absorbance of the 0-acetylserine sulfhydry1ase:Oacetylserine solution could be reversed by titration with NazS (data not shown). In an 0-acetylserine-saturated solution, addition of 4.0 p~ NazS regenerated the original spectrum with the absorbance maxima at 412 nm, probably by its reaction with the intermediate to transfer the acetyl group of O-acetylserine to sulfide, forming the L-cysteine product (Cook and Wedding, 1976). The observed changes in 0-acetylserine sulfhydrylase absorbance spectra upon addition of 0-acetylserine and reversion with addition of NazS illustrate the involvement of the pyridoxal 5'-phosphate cofactor in the catalytic activity of . Absorption spectrum of 0-acetylserine sulfhydrylase isoenzyme A. The spectrum was measured on 34 pg/ml protein in 10 m~ Tris, pH 8.0. The absorption spectra of isoenzymes B and C were similar. Note the absorption maxima at 278 and 412 nm, due to the presence of aromatic amino acids in the protein and the presence of bound pyridoxal 5'-phosphate, respectively. this enzyme. The spectral shift to 460 and 330 nm also spontaneously reverted to the original spectrum after several hours of storage, probably due to the instability of 0-acetylserine at neutral pH (Kredich and Tomkins, 1966;Becker et al., 1969).

0-Acetylserine Sulfiydrylase Isoenzymes from D. innoxia
Effects of Temperature and p H on 0-Acetylserine Sulfhydrylase Activity-The three 0-acetylserine sulfhydrylase isoenzymes responded very similarly to changes in reaction temperature and pH. They were active over a broad range of temperature (Fig 6A). Activity generally increased from 20 to 30 "C and leveled off between 30 and 40 "C. Maximum activity was observed at temperatures between 42 and 58 "C. 0-Acetylserine sulfhydrylase activity was highly dependent on reaction pH. The enzymes were only active in the range of 7 to 8, with optimal activity of all isoenzymes at pH 7.6. Even slight changes in pH above or below this value dramatically reduced activity. For this reason, pH was carefully monitored at all concentrations of 0-acetylserine and NazS used in enzyme assays. The purified 0-acetylserine sulfhydrylase proteins were irreversibly inactivated in buffers a t pH below 6.8 and during isoelectric focusing, despite attempts to regain activity by addition of cofactor and pH adjustment. Similar inactivation has been observed in the 0-acetylserine sulfhydrylase from other organisms (Ikegami et al., 1988;Cook et al., 1992). Above pH 8.0, the 0-acetylserine substrate undergoes a n 0-to N-shift, and is no longer a substrate for the enzyme (Murakoshi et al., Ikegami et al., 1988). By spectral analysis, Cook et al. (1992) have shown that at pH greater than 8.1 the €-amino group of the active site lysine from bacterial 0-acetylserine sulfhydrylase becomes deprotonated resulting in deacetylase activity. This generates pyruvate and ammonia from O-acetylserine, instead of cysteine.
Substrate Saturation Curves-The in vitro catalytic ability of the three 0-acetylserine sulfhydrylase forms was similar  (Table 11). Results are illustrated for isoenzyme C only (Figs. 7 and 8).
0-Acetylserine-The saturation curves for 0-acetylserine were sigmoid (Fig. 7b), suggesting the enzymes are allosteric and exhibit positive cooperativity with respect to this substrate. Note that all three equations used to fit the sigmoid portion of the 0-acetylserine saturation curve have similar fits, indicated by the similar mean square error (M.S.E.) and coefficient of determination ( R 2 ) values in Table 11. The MMF model is a generalization of the Hill equation in that it allows for a non-zero intercept. Since we have no information on the shape of the curve very close to zero, it may be incorrect to assume the curve is strictly sigmoidal from zero to the observed data points. Allowing a non-zero intercept allows the model to fit more closely to the data and not extrapolate from the data down to zero. Since the Hill equation forces the curve down through zero, the So.s tends to be reached sooner (smaller value) than using the MMF equation. The logistic equation, as a very general growth equation, produces estimates of V, , , and So.5 that are comparable to the other two models for the 0acetylserine substrate. The data obtained from 0-acetylserine saturation experiments of isoenzymes A and C provided better fits to the three models than isoenzyme B. The data set for isoenzyme B contained a great deal of scatter at 0-acetylserine concentrations above 10 mM. This is reflected in the MSE and R2 values for this isoenzyme in Table 11.
Based on our findings for subunit size and cofactor presence in the 0-acetylserine sulfhydrylase proteins, we expect there to be a t least two active sites in these enzymes, with 1 molecule of pyridoxal 5'-phosphate per site. The values of n, predicted from the Hill and MMF equations, are consistent with this hypothesis. The Hill equation characteristically underestimates the amount of cooperativity, and the MMF equation appeared to overestimate this parameter.
In initial experiments, 0-acetylserine sulfhydrylase activity appeared to be inhibited by 0-acetylserine concentrations above 25 mM (Fig 7a). Cook and Wedding (1976) reported a similar finding for 0-acetylserine from S. typhimurium. Measurements of reaction pH in these assays demonstrated that the high concentrations of 0-acetylserine in these assays were reducing the reaction pH to below the narrow optimum range for 0-acetylserine sulfhydrylase activity (Fig. 6B ). To compensate for this, the pH of the 100 mM Tris reaction buffer was increased incrementally by 0.2 increment, so that addition of the Na2S and 0-acetylserine substrates resulted in a final reaction pH of 7.6. When assays were conducted under these conditions, no inhibition was observed.
Sodium Sulfide-The early portion of the sodium sulfide saturation curves, up to 200 p~, were sigmoid ( Fig. 86) but the curves did not smoothly approach an asymptote. Instead they peaked and decreased abruptly. The addition of an inhibition curve to the sigmoid saturation curve made these curves more difficult to fit. The logistic model was able to fit the early part of the curve, and appears to provide the best estimates of V,,, and So.5 in the presence of the inhibition curve. The Hill and MMF equations did not fit the early part of the curve as well and there is more uncertainty in the estimates of the asymptote (V,, ) and So.s using these models. Because of the inhibition portion of the curve, it is difflcult to extrapolate physiological conclusions about the three isoenzymes based on these in vitro results. The estimates of n from the Na2S data were between 1 and 3, which is consistent with the results of the 0-acetylserine saturation curves, and our physical data on these proteins. The Na2S substrate was inhibitory to all three enzymes at concentrations above 200 p~. Similar inhibition has been described in 0-acetylserine sulfhydrylase from Phaseolus sp. (Bertagnolli and Wedding, 1969) and S. typhimurium (Cook and Wedding, 1976).

DISCUSSION
Multiple forms of 0-acetylserine sulfhydrylase have been identified in plants using nondenaturing polyacrylamide gel electrophoresis (Bertagnolli and Wedding, 1977) and anion exchange chromatography (Ikegami et al., 1987(Ikegami et al., ,1988Nakamura and Tamura, 1989;Lunn et al., 1990;Rolland et al., 1992). The number of isoenzymes identified from different plant species ranges from one to five. Most plants probably contain multiple forms of this enzyme, and some of the early studies of plant 0-acetylserine sulfhydrylase enzymes were probably accomplished using mixtures of isoenzymes. We separated and purified three forms of the enzyme from D. innoxia with similar in vitro catalytic abilities but different physical properties. These three isoenzymes probably perform different functions in the plant cell.
In the bacterium, S. typhimurium, 0-acetylserine sulfhydrylase activity is present as a free form and in association with serine transacetylase as a 309-kDa multienzyme complex comprised of 2 molecules of 0-acetylserine sulfhydrylase and 1 molecule of serine transacetylase (Kredich and Tomkins, 1966;Becker et al., 1969;Kredich et al., 1969). We found no evidence of a similar complex in non-photosynthetic D. innoxia cell cultures, either by separation of crude protein extracts on anion exchange HPLC, or gel filtration chromatography. Similarly, Bertagnolli and Wedding (1977) found no evidence for such a complex in green tissue from two Phaseolus sp. using activity assays for both 0-acetylserine sulfhydrylase and serine transacetylase. In contrast, Droux et al. (1992) recently identified that 3-5% of the 0-acetylserine sulfhydrylase activity in Spinacia oleracea chloroplasts eluted as a 310-kDa complex that also contained serine transacetylase activity. If a similar complex exists in D. innoxia chloroplasts, it is possible that it is not formed in the non-photosynthetic plastids of the cell cultures used in this study. We are currently investigating the subcellular location of 0-acetylserine sulfhydrylase isoenzymes from cell cultures and photosynthetic and non-photosynthetic tissues of D. innoxia plants.
Most investigators have assigned Michaelis-Menten kinetics to 0-acetylserine sulfhydrylase enzymes Masada et al., 1975;Ascaiio and Nicholas, 1977;Leon et al., 1987;Murakoshi et al., 1985;Ikegami et al., 1987Ikegami et al., , 1988Nakamura and Tamura, 1989;Drow et al., 1992). With multiple data points and replicated experiments we have shown that this is clearly not the case for the three 0-acetylserine sulfhydrylase proteins from D. innoxia. Our work is in agreement with the results of Bertagnolli and Wedding (19771, who found that 0-acetylserine sulfhydrylase from Phaseolus sp. exhibited positive cooperativity at lower concentrations of sulfide. The D. innoxia isoenzymes exhibit positive cooperativity with both the 0-acetylserine and sulfide substrates. The 0-acetylserine sulfhydrylases from D. innoxia were inhibited by the sulfide substrate. However, in contrast to the S. typhimurium O-acetylserine sulfhydrylase (Cook and Wedding, 1976;Bertagnolli and Wedding, 1977), they were not inhibited by 0-acetylserine as long as the reaction pH was maintained in the narrow, optimal range required for activity.
In the plant cell, cysteine synthesis is sensitively regulated To avoid the substrate inhibition portion of the curve (Fig. 8), analysis was restricted to sodium sulfide concentrations between 0 and 200 w for isoenzymes A and C, and between 0 and 150 p~ for isoenzyme B. illustrates results from reactions that were not corrected for pH. b, expanded illustration of the early points in the curve, illustrating the sigmoid form of the curve and comparative fits by three non-linear models as described under "Experimental Procedures." through 0-acetylserine sulfhydrylase activity. This activity is specific for and highly responsive to the enzyme substrates. 0-Acetylserine and sulfide affect the enzyme in a positive manner at low concentrations, and sulfide inhibits activity as concentration of this substrate increases. Previous studies have shown substrate specificity for both the alanyl moiety and the sulfide donor in plant 0-acetylserine sulfhydrylases (Murakoshi et al., 1985;Ikegami et al., 1988). Free sulfide (in the form of Na'S) is the only sulfide source that has been shown to be a substrate for 0-acetylserine in vitro, and whether a sulfide carrier is involved in the in vivo synthesis of cysteine is still unknown (Schmidt and Jager, 1992).

0-Acetylserine
The three isoenzymes in our in vitro analysis were most highly active at temperatures higher than physiologically normal. Glutathione concentrations increase in plants exposed to heat shock (Nieto-Sotelo and Ho, 1986), suggesting a n increase in biosynthesis of this tripeptide. While cysteine levels do not increase, this amino acid is a required precursor for glutathione and it is reasonable that cysteine synthesis would increase in response to heat shock. The observed high activity of the 0acetylserine sulfhydrylase isoenzymes at temperatures between 42 and 58 "C suggests a mechanism for increased cysteine synthesis in response to elevated temperatures. at each substrate concentration. b, expanded illustration of the early points in the curve, illustrating the sigmoid form of the curve and comparative fits by three non-linear models as described under "Experimental Procedures." 0-Acetylserine sulfhydrylase enzymes play several metabolic roles in plants. They are responsible for the final step in the assimilation of inorganic sulfur into a n organic molecule and therefore play an essential role in primary plant metabolism. In addition, 0-acetylserine sulfhydrylases provide one of the necessary precursors (cysteine) for the synthesis of glutathione and the metal-binding polypeptides involved in cellular response to oxidative stress induced by toxic metals, heat, radiation, and other perturbations. We and others (Lunn et al., 1990;Rolland et al., 1992)' have demonstrated that in plants, different 0-acetylserine sulfhydrylase isoenzymes are predominant in the cytosol, chloroplasts, and mitochondria. We have initiated studies to examine the involvement of the 0-acetylserine sulfhydrylase isoenzymes in different aspects of plant cell metabolism, to determine whether the O-acetylserine sulfhydrylase proteins localized in different regions of the plant cell play different roles in sulfur assimilation and transport, protein synthesis, glutathione accumulation and responses to external abiotic stresses. With the purified O-acetylserine sulfhydrylase proteins described here, antisera against these proteins, and the genes encoding 0-acetylserine C. R. Kuske and P. J. Jackson, unpublished results 0-Acetylserine Sulfhydrylase Isoenzymes from D. innoxia sulfhydrylase in D. innoxia,2 we can begin to address these questions.