Serine hydroxymethyltransferase: origin of substrate specificity

Roberto Contestabile, Sebastiana Angelaccio, Francesco Bossa, H. Tonie Wright, Neel Scarsdale, Galina Kazanina, and Verne Schirch. Biochemistry 2000 3...
0 downloads 0 Views 1MB Size
Biochemistry 1992, 31, 155-162

Tsujita, T., & Brockman, H. L. (1987) Biochemistry 26, 8423-8429. Tsujita, T., Smaby, J. M., & Brockman, H. L. (1987) Biochemistry 26, 8430-8434. Tsujita, T., Muderhwa, J. M., & Brockman, H. L. (1989) J . Biol. Chem. 264, 8612-8618.

155

Verger, R., & Pieroni, G. (1986) in Lipids and Membranes: Past, Present and Future (Op den Kamp, J. A. F., Roelofsen, B., & Wirtz, K. W. A., Eds.) pp 153-170, Elsevier Science Publishers, Amsterdam. Winkler, F. K., D’Arcy, A., & Hunziker, W. (1990) Nature 343, 771-774.

Serine Hydroxymethyltransferase: Origin of Substrate Specificity7 Sebastiana Angelaccio,* Stefan0 Pascarella,* Elena Fattori,t Francesco Bossa,* William Strong,§ and Verne Schirch*v§ Department of Biochemistry and Molecular Biophysics, Virginia Commonwealth University, Richmond, Virginia 23298, and Dipartimento di Scienze Biochimiche and Centro di Biologia Molecolare del Consiglio Nazionale delle Ricerche, Universitd La Sapienza, 00185 Roma, Italy Received July I , 1991; Revised Manuscript Received September 10, 1991

ABSTRACT: All forms of serine hydroxymethyltransferase, for which a primary structure is known, have five threonine residues near the active-site lysyl residue (K229) that forms the internal aldimine with pyridoxal phosphate. For Escherichia coli serine hydroxymethyltransferase each of these threonine residues has been changed to an alanine residue. The resulting five mutant enzymes were purified and characterized with respect to kinetic and spectral properties. The mutant enzymes T224A and T227A showed no significant changes in kinetic and spectral properties compared to the wild-type enzyme. The T225A and T230A enzymes exhibited differences in K,,, and kat values but exhibited the same spectral properties as the wild-type enzyme. The four threonine residues at positions 224,225,227, and 230 do not play a critical role in the mechanism of the enzyme. The T226A enzyme had nearly normal affinity for substrates and coenzymes but had only 3% of the catalytic activity of the wild-type enzyme. The spectrum of the T226A enzyme in the presence of amino acid substrates showed a large absorption maximum at 343 nm with only a small absorption band at 425 nm, unlike the wild-type enzyme whose enzymesubstrate complexes absorb at 425 nm. Rapid reaction studies showed that when amino acid substrates and substrate analogues were added to the T226A enzyme, the internal aldimine absorbing at 422 nm was rapidly converted to a complex absorbing at 343 nm in a second-order process. This was followed by a very slow first-order formation of a complex absorbing at 425 nm. Variation of the initial rapid second-order process as a function of pH suggested that the anionic form of the amino acid forms the first complex with the enzyme. The results are interpreted as being due to the rapid formation of a gem-diamine complex between amino acids and T226A enzyme with a ratedetermining formation of the external aldimine. This suggests that Thr-226 plays an important role in converting the gem-diamine complex to the external aldimine complex. Variation of the kinetic constants with amino acid structure suggests that the T226A enzyme distinguishes between substrates and substrate analogues in the formation of the gem-diamine complex.

S e r i n e hydroxymethyltransferase (SHMT) catalyzes the conversion of serine and tetrahydrofolate (H,folate) to glycine and 5,10-methyleneH4folate. This reaction is present in a wide variety of cells and is the major source of one-carbon groups required in the biosynthesis of methionine, choline, thymidylate, and purines (Schirch, 1982). We have previously purified and determined the primary structure of cytosolic and mitochondrial isoenzymes from rabbit liver (Martini et al., 1987, 1989). SHMT has also been purified and characterized from expression of the Escherichia coli cloned glyA gene (Plamann et al., 1983; Schirch et al., 1985). SHMT activity is dependent on the two coenzymes pyridoxal-P and H4folate. Pyridoxal-P is covalently attached at the active site and serves as a spectrophotometric probe in determining the structure of en‘This work was supported in part by Grant GM 28143 from the National Institutes of Health and a grant from the Istituto PasteurFondazione Cenci Bolognetti. * To whom correspondenceshould be addressed. Universitd La Sapienza. $ Virginia Commonwealth University.

*

zyme-substrate intermediates in the reaction pathway (Schirch, 1982). As with all pyridoxal-P enzymes, the site of covalent attachment is an internal aldimine between the 4’-aldehyde group on the coenzyme and an e-amino group of a lysyl residue (Davis and Metzler, 1972). Reduction of this external aldimine converts it to a stable secondary amine, which permits isolation of a peptide from proteolytic digests containing the bound pyridoxal-P (Bossa, et al., 1976). These active site peptides have been isolated and sequenced from numerous pyridoxal-P containing enzymes (Vaaler and Snell, 1989; Tanizawa, et al., 1989). The three forms of SHMT which we have studied all contain the nine-residue conserved sequence V-V-T-T-T-T-H-K(Pyr)-T (Martini et al., 1989). The active-site nonapeptide from SHMT is unusual in that five Thr residues have been conserved in the E. coli enzyme and the mammalian isoenzymes in rabbit liver. This suggests that these Thr residues have some functional role in this en-



Abbreviations: SHMT, serine hydroxymethyltransferase;H,folate, tetrahydrofolate; pyridoxal-P, pyridoxal 5’-phosphate.

0006-2960/92/043 1-155$03.00/0 0 1992 American Chemical Society

156 Biochemistry, Vol. 31, No. 1, 1992 zyme. There are at least two functional roles these residues could have in the mechanism of this reaction. First, they could serve as hydrogen-bond donors or acceptors to either of the two coenzymes or to the substrate amino acids glycine and serine. Serving as hydrogen donors or acceptors, the Thr residues could play a role in determining both reaction and substrate specificity. Second, they could serve as acceptors and donors of protons between various intermediates in the reaction pathway. Including known and proposed mechanistic steps in this reaction, there may be as many as nine different steps in which protons are transferred between some group on the enzyme and either the substrate or both coenzymes. SHMT utilizes a wide variety of 3-hydroxyamino acid substrates and catalyzes numerous side reactions, which include decarboxylation, transamination, and racemization (Shostak & Schirch, 1988). Previously, we have provided evidence that a conformational change that occurs on substrate binding to SHMT controls reaction specificity (Schirch et al., 1991). What has not been elucidated is at what step in the reaction pathway the enzyme distinguishes between substrate amino acids serine and glycine and nonsubstrate amino acids such as L-alanine and threonine. The purpose of the study reported here was to determine the function of each of the five Thr residues in E . coli SHMT. Each Thr was changed to an Ala residue by site-directed mutagenesis, and the five resulting mutant proteins were purified and characterized with respect to affinity of coenzymes, substrates, and substrates analogues. Also, kinetic properties in the aldol cleavage of substrates and substrate analogues and ability to catalyze the transamination of D- and L-alanine were studied, and rapid reaction studies to determine the rate of interconversion of enzyme-substrate complexes were performed. One of these mutant proteins provides insight into how the enzyme controls which amino acids are accepted as substrates and which amino acids are rejected. EXPERIMENTAL PROCEDURES Materials. Amino acid substrates, NADP+, NADH, alcohol dehydrogenase (yeast), and pyridoxal-P were obtained from Sigma. (6RS)-H4folate was purchased from Fluka. C1-Tetrahydrofolatesynthase was purified from rabbit liver as previously described ( W a r et al., 1985). Removal of bound pyridoxal-P to form aposerine hydroxymethyltransferase was achieved by incubation of the enzyme with L-cysteine in high salt (Schirch et al., 1973). Site-Directed Mutagenesis. Threonine residues at amino acid positions 224 and 227 are coded for by ACT and threonine residues at the amino acid positions 225, 226, and 230 are coded for by the sequence ACC in the glyA gene (Plamann et al., 1983). These five threonine residues were changed individually to alanine residues by making five oligonucleotides that changed these codons to either GCT or GCC, which code for alanine. Thr-226 was also converted to a Ser residue. The oligonucleotide mutagenesis kit from Amersham was used to obtain each mutant of the glyA gene in a single-stranded M 13mp9 clone. After the sequence of the mutant gene in the M13mp9 insert was verified, the mutant glyA gene was transferred to the plasmid pBR322 and used to transform E . coli strain GS245 as previously described (Hopkins & Schirch, 1986). The mutant form of each enzyme was purified using the same procedure described for the wild-type enzyme (Schirch et al., 1985). For mutant proteins T224A, T225A, T226A, T226S, and T227A, the substitution of either alanine or serine for threonine was verified by amino acid sequence analysis as described by Barra et al. (1991). For mutant protein T230A the mutation was verified by sequence analysis

Angelaccio et al. of the double-stranded expression plasmid in the GS245 cells. Kinetic Studies. Kinetic studies were performed using several different substrates. When L-serine was used as the changing fmed substrate, H4folate, at concentrations of 10-100 pM, was used as the variable substrate. C,-Tetrahydrofolate synthase was used as the coupling enzyme, and the reaction rate was determined from the increase in absorbance at 340 nm due to the reduction of NADP' (Schirch & Peterson, 1980). K,,, values were determined from double-reciprocal plots. Both allothreonine and threonine were also used as substrates. Each of these reactions was followed by determining the rate of reduction of the product acetaldehyde by alcohol dehydrogenase and NADH at 340 nm (Schirch & Peterson, 1980). The rates of transamination of both L- and D-alanine were determined from the rate of decrease in absorbance at 425 nm in solutions of enzyme (1 mg/mL), 100 mM D- or L-alanine, and 50 mM potassium phosphate buffer at pH 7.6 at 37 OC (Shostak & Schirch, 1988). The affinity of each mutant apoenzyme for pyridoxal-P was determined by incubating 0.1 pM solutions of the apoenzyme with concentrations of 0.05-5 pM solutions of pyridoxal-P at room temperature in 20 mM potassium N,N-bis[2-hydroxyethyl]-2-aminoethanesulfonate,pH 7.0, for 3 h. Each solution was assayed and the concentration of holoenzyme determined from the fraction of activity of these solutions compared to the activity of the apoenzyme incubated with 50 pM pyridoxal-P. The concentration of remaining apoenzyme in each solution was the difference between the concentration of the original total apoenzyme and the concentration of holoenzyme determined from the activity measurements. The concentration of unbound pyridoxal-P was determined from the difference in concentrations of pyridoxal-P added to the apoenzyme solutions and the holoenzyme. Kd values were determined from the equation Kd = [apoenzyme][pyridoxalP] / [ holoenzyme]. Rapid Reaction Studies. Stopped-flow absorbance measurements were performed on a Kinetics Instruments spectrophotometer. Temperatures were held at either 8,20, or 30 OC by a circulating water bath. Traces were recorded on a MacIntosh IIc using the software provided by Kinetic Instruments. Each study was an average of 4-6 traces. The curves for absorbance versus time were curve-fit by either a single- or double-exponential algorithm. Rate constants and amplitudes for each individual reaction varied less than 10% from the average values for each reaction. The absorbance changes observed in the stopped-flow studies were in agreement with the predicted absorbance changes from equilibrium spectral studies. For studies performed at either 8 or 20 OC, the enzyme concentration in the reaction vessel was 2.0 mg/mL. This gives a subunit concentration of 43 pM. For the reactions done at 30 "C the concentration of enzyme in the reaction vessel was 2.5 mg/mL. The buffer used for both enzyme and substrate solutions was either 50 mM potassium N,N-bis[2-hydroxyethyl] -2-aminoethanesulfonate, pH 7.0, or 50 mM potassium phosphate at the indicated pH. The concentration of the anionic forms of amino acids were calculated using the Henderson-Hasselbalch equation and pK, values of 9.6,9.2, 10.4, and 9.7 for the amino groups of glycine, L-serine, m-allothreonine, and L-alanine, respectively. Spectral Studies. Ultraviolet-visible spectra were recorded on a Cary 210 spectrophotometer and circular dichroism spectra were obtained with a Jasco 500 spectropolarimeter. RESULTS Kinetic Constants. Each of the purified mutant enzymes containing an alanine in place of a threonine was tested for

Biochemistry, Vol. 31, No. I, 1992 157

Serine Hydroxymethyltransferase

Table I: Kinetic Constants for Wild TvDe and Five Mutant Forms of E . coli Serine Hvdroxymethvltransferase form of serine hydroxymethyltransferase substrate kinetic constant wild type T224A T225A T226A T227A T230A threonine Km (mM) 12 10 7 20 8 20 k,, ( m i d ) 2.2 2.9 6