Bit~'himica et Biophysica Acla, 1121 (1992) 234-238 ~' _"~1992 Elsevier Science Publishers B.V. All rights reserved 016%4838/92/$05.00
234
BBAPRO 34205
Identification of an essential cysteine in the reaction catalyzed by aspartate-fl-semialdehyde dehydrogenase from Escherichia coil William E. Karsten ' and Ronald E. Viola Departmcn, of Chemist~'. The Unicersit)"of Akron. Akron. OH (USA)
(Received 7 October 1~91) (Revised manuscript received 6 January 1992)
Key words: Enzymemechanism: Oligonucleotide-directed mutagenesis: Enzyme kinetics: Essential cysteine: pH profile The enzyme L-aspartate-/3-semialdehyde dehydrogenase from Eschenchia coil has been studied by oligonucleotide-directed mutagenesis. The focus of this invcstigation was to examine the role of a cysteine residue that had been previously identified by chemical modification with an active sitc directed reagent (Biclimann et al. (1980) Eur. J. Biochem. 104, 59-64). Substitution of this cysteinc at position I35 with an alanine results in complete loss of enzyme activity. However, changing this cysteine to a scrine yields a mutant enzyme with a maximum velocity that is 0.3% that of the native enzyme. This C135S mutant has retained essentially the same affinity for substrates as the native enzyme, and the same overall conformation as reflected in identical behavior on gel electrophoresis and in identical fluorescence spectra. The pH profile of the native enzyme shows a loss in catalytic activity upon protonation of a group with a pK., value of 7.7. The same activity loss is observed at this pH with the ~rine-135 mutant, despite the differences in thc pK a values for a cysteine sulfhydryl and a serine hydroxyl group that have been measured in model compounds. This observed pK~ value may reflect the protonation of an auxmary catalyst that enhances the reactivity of the active site cysteinc nucleophile in the native aspartate-/3-semialdehyde dehydrogenase.
Introduction
L-Aspartate-/3-semialdehyde dehydrogenase (E.C.I. 2.1.11) is a dimeric enzyme composed of identical subunits. The DNA sequence [1] predicts a subunit composed of 367 amino acids with an overall molecular weight of 39950. Starting from L-aspartate, the amino acid biosynthetic pathway in Escherichia coil leads to the formation of L-lysine, L-isoleucine, L-methionine and L-threonine. The first and third steps in this pathway are catalyzed by the bifunctionai enzymes aspartokinase - homoserine dehydrogenase I and II. A third monofunctional enzyme, aspartokinase Ill, also catalyzes the initial reaction in the pathway, and each of these enzymes are differentially regulated to control
the levels of each end product amino acid [2]. L-Aspartate-/~-semialdehyde dehydrogenase (ASA-DH)catalyzes the intervening branch point reaction between the aspartokinase and the homoserine dehydrogenase reactions, with one of these branches leading to the production of L-lysine and the other leading to Lmethionine, L-threonine and 1.-isoleucine. In the biosynthetic direction ASA-DH catalyzes the formation of L-aspartate-/3-semialdehyde (ASA) by the reduetive dephosphorylation of L-/3-aspartyl-phosphate (BAP) utilizing NADPH: COO ~ H3N_CH+NADPH+H + i
C(O)OPOi i Present address: Department of Microbiology and Immunology, Texas College of Osteopathic Medicine, Fort Worth, TX, USA. Abbreviations: ASA, aspartate-~-semialdehyde: ASA-DH, asparlate/3-semialdehyde dehydrogenase, BAP, /3-aspartylphosphate: DTNB, 5,5'-dithiobis42-nitrobenzoic acid): DTI', dithiothreilol; NEM, Nethylmaleimide. Correspondence: R.E. Viola, Department of Chemistry, The University of Akron, Akron, OH 44325-3601, USA.
,43-asparlyl phosphate
COO
-
÷H~N_CH+H2PO4-+NADP ÷ "
~{C-O i -asparlatc-/3. -scmialdehyde
The ASA-DH reaction is related to the reaction catalyzed by glyceraldehyde-3-phosphate dehydrogenase and similar chemical mechanisms have been proposed for each enzyme [3-5]. The suggested mechanism involves the formation of a thioester intermediate resulting from attack of a cysteine thiolate on the
235 substrate carbonyl group, followed by hydride transfer to NADP. Subsequent attack on the thioester intermediate by an oxygen anion of bound inorganic phosphate leads to expulsion of the cysteine thiolate and formation of the phosphorylated product. Chemical modification studies with DTNB [31 and N-ethylmaleimide [5] have indicated an essential role for a cysteine thiol in the ASA-DH reaction_.,. Substrate protection studies against N-ethylmaleimide inactivation performed with ASA have suggested that the essential cysteine is lo-cated in or near the enzyme active site. In addition, pH studies have indicated a role for a neutral acid group that must be ionized for enzymatic activity. The p K a value determined for this group is similar t-q that for the enzymatic group that leads to loss of enzyme activity when modified by N-ethylmaleimide [5]. L-2-Amino4-oxo-5-chloropentanoic acid, an alkylating substrate analogue of ASA, has been shown to irreversibly inactivate ASA-DH. An isolated peptide containing the modified amino acid, tentatively identified as a histidine, was reported to have the amino acid sequence Phe-Val-Gly-Gly-Asp-His-Thr-Val-Ser [6]. Haziza [1] determined the nucleotide sequence of the asd gene of E. coli that codes for ASA-DH and found a corresponding sequence encoding for amino acid residues 130-138 of their deduced sequence. However, they found a substitution of asparagine for aspartic acid and cysteine for histidine at residues 134 and 135. These results are consistent with the cysteine located at position 135 of the amino acid sequence being the most likely candidate for the active site cysteine. Utilizing site-directed mutagenesis we have constructed two mutants of ASA-DH in order to further clarify the role of cysteine 135 in the chemical mechanism of the enzyme. We report here on the preparation, activity and characterization of these mutant enzymes, and on the implications of these results on the chemical mechanism of ASA-DH.
Experimental procedures Materials. Restrict;on enzymes were purchased from Promega or U.S. Biochemical Corporation. T4 DNA ligase, MI3K07 bacteriophage, bacterial strain JM109 and the pGEM-7Zf( + ) plasmid vector were purchased from Promega. The strain of E. coil harboring the plasmid pOP126 containing the asd gene [7,8] was a gift from Dr. Spencer Shames. The E. coli strains RZ1032 and HBI01 were obtained from Dr. J.J. Villafranca. The deletion mutant devoid of ASA-DH activity, E. coli KI2 G6MD3 [7], was a gift from Dr. Jack Preiss. Unpurified oligonucleotides were purchased from Operon Technologies. Sequencing of DNA was performed by the dideoxynueleotide chain termination method using the Sequenase DNA sequencing kit purchased from U.S. Biochemical Corporation.
Bacterial growth and media. A standard enriched media was used for growth of all bacteria and contained Bacto-tryptone (8 g/l), yeast extract (5 g/l) and sodium chloride (2.5 g/l). E. coli strain K12 G6MD3 requires diaminopimelate, and was grown on media containing 50 # g / m l diaminopimelate. Ampicillin was included at a concentration of 100 # g / m l where required. Subcloning and mutagenesis. The plasmid pOP126 contains the asd gene on an approx. 1600 base pair fragment inserted into the Bam H I site of pBR322. This BamH ! fragment from pOP126 was subcloned into the Bam H ! site of the pGEM-7Zf(+ ) plasmid vector to conduct the mutagenesis studies. The Kunkel site-directed mutagenesis method [9] utilizing uracilcontaining ssDNA was performed on the pGEM subclone of the asd gene. The cysteine-135 codon (TGT) was replaced with either alanine (GCT) or serine (TCG) by using 5'-GCTI'ACGGTAGCGTTACCGCC-3' as the primer for the alanine mutation (base changes underlined), and the corresponding 21-base oligonucleotide for the serine mutation. The substitution of each of these two nucleotides for the respective wild type bases also eliminates an Rsal restriction enzyme site that overlaps the TGT cysteine codon in the native DNA. Rsal restriction enzyme digests to detect the presence or absence of this restriction enzyme site were used to carry out the initial screening of plasmids for the presence of mutant DNA. The amino acid substitutions were verified by determination of the nucleotide sequence in the region of the gene that codes for position-135. Sequencing was initiated within several hundred bases of this position by utilizing synthetic oligonueleotides as primers. Characterization of enzymes. The C135S and wild type enzymes were purified by the method of Karsten and Viola [5]. Protein concentrations were determined by the method of Bradford [15]. Enzyme activity was routinely assayed in the phosphorylating direction at 25°C in 1 ml cuvettes containing varying concentrations of substrates and 0.5 mM DTT. For the variation of kinetic parameters as a function of pH the buffers used were Mes (pH 6.0-6.5), Hepes (pH 7.0-8.0), and Ches or Taps (pH 8.5-9.5) at a concentration of 100 mM. Enzyme assays were performed on a Gilford 260 recording spectrophotometer equipped with thermospacers and connected to a circulating water bath to maintain a constant temperature. The fluorescence spectra of the native and thc mutant enzymes were examined on a Shimadzu RF-5000U fluorescence spectrometer in 50 mM Hepes (pH 7.0), with 0.5 mM DTT. The enzyme solutions were excited at 280 nm and the emission spectra were recorded from 300 to 400 nm. Chemical modification studies with N-ethylmaleimide were carried out in 100 mM Hepes (pH 7.0). From 7-20 ~g of enzyme (either
236 native or C135S mutant) was incubated with 0.2 mM NEM. Aliquots were removed at various times and examined for enzymatic activity in an assay mixture in 150 mM Ches (pH 9.0) containing 0.1 mM ASA, 0.2 mM NADP, 25 mM potassium phosphate and 0.5 mM DTI'. Data analysis. The kinetic data were analyzed by using BASIC versions of the kinetics programs of Cleland [10]. Initial velocity data giving intersecting patterns were fitted to Eqn. 1 where c = initial velocity; V= maximum velocity; A or B = substrate concentrations; K = Michaelis constants. c=
VAB KiaK~ + Ka B + KbA + AB
(I)
For the pH profiles kinetic parameters were determined at each pH by varying the concentration of ASA at saturating levels of the other two substrates and fitting the initial velocity data to Eqn. 2. For the serine mutant, the V / K values determined at each pH were fitted to Eqn. 3, which allows for activity decreasing in the acid region of the pH profile, and where C is the maximum V / K value and K~ is the acid dissociation constant. For the native enzyme, the V / K values were fitted to Eqn. 4, which allows for a decrease in activity in both the acid and basic regions. VA
c=
K+A
(2)
log V / K = log{C/( I + [H " I/K..,)}
(3)
log V / K = log{C,/( I -)-[H" I/K., + K h / I H ÷ l)}
(4)
Results
Construction and purification of mutants Mutants of ASA-DH were constructed by oligonucleotide-directed mutagenesis to examine the role of a putative active site cysteine in the mechanism of the enzyme-catalyzed reaction. Initial screening for the presence of the Ala-135 or Ser-135 mutant was accomplished by Rsal restriction enzyme digests of isolated plasmids. The loss of the Rsal restriction enzyme site that overlaps the wild type cysteine codon was taken as evidence for the presence of the desired mutation. Plasmids producing the appropriate restriction map were then sequenced through the region coding for the amino acid at position 135, by using the dideoxy chain termination method, to verify the identity of the amino acid produced at that position. Plasmids that were identified to contain the desired mutation were used to transform the deletion mutant G6MD3 E. coli strain that is essentially devoid of ASA-DH activity. Cultures of these transformed cells were grown overnight, and then disrupted by sonic oscillation. The crude extracts from these transformed cells were assayed for the
TABLE I Acticity of ASA-DH mutants
Enzyme"
Wild type C 135A C135S
Activity h pH 7
(%)
pH 9
(%)
2.1 0.01)06 (I.0049
(100) (0.(13) (0.2)
15.7 0.00016 0.052
(100) (0.001) (0.3)
Extracted from the G6MD3 strain of E. coli containing the pGEM plasmid with an insert of either the wild type or a mutated asd gene. h Assay conditions: 200 mM Hepes (pH 7.0) or 200 mM Ches (pH 9.0), 0.2-[I.5 mM NADP, 0.15 mM ASA, 50 mM Pi and 0.5 mM DTI'. Enzyme activity in crude cell extracts, expressed in 0,tool/rain per mg of total protein, or as a percentage of the wild type activity at that pH.
acquisition of ASA-DH activity. The Ala-135 mutant displays some very low ASA-DH activity in crude extracts at pH 7, however, the level of activity is virtually identical to that observed in the G6MD3 strain that does not contain the plasmid. Essentially no enzyme activity is observed in the Ala-135 mutant strain at pH 9 (Table !), which is the pH optimum for wild type enzyme activity. Also, partial purification of this C135A mutant protein resulted in the loss of this background activity. The Ser-135 mutant enzyme showed activity at both pH 7 and 9, but the activity that was measured was well below the wild type levels that is typically observed in these crude extracts. Samples of the crude extracts obtained from each of these mutant strains were run on SDS-PAGE and stained with Coumassie blue. In each case a protein band comprising approx. 15% of the total protein was present in the sample at a position identical with that of the purified wild type ASA-DH. The Ser-135 mutant enzyme was purified from these crude extracts in order to more fully characterize the activity of this enzyme. The mutant enzyme behaved in an identical fashion to the wild type enzyme throughout the purification. The purified Ser-135 mutant was determined to be > 95% pure by Coumassie blue stained protein gels.
Characterization of the Ser-135 mutant enzyme The fluorescence spectra of the native and the CI35S enzymes were examined from 300 to 400 nm. The maxima for each enzyme was observed at 320 rim, and the spectra of each were superimposeable (data not shown). Reaction of N-ethylmaleimide (NEM)with the native enzyme results in a rapid and complete loss of enzyme activity with a corresponding modification of a single cysteine residue [5]. When the serine-135 mutant is treated with NEM there is only about a 15% decrease in activity after incubation with the reagent for 15 rain. Treatment of the native enzyme under these
237 T A B L E II
Comparison of the kinetic parameters between wild type and Ser-135 mutant of ASA-DH Parameter
Wild type
K m (mM) ~ ASA NADP P+
C135S
Ratio w.t./mutant
0.057 (0.{172) b 0.038 (0.017) 1.86 (0.33)
0.146 (0.02b) 0.018 (0.005) 0.45 (0.1)
4.07. l i p
3.24-103
0.4 2.1
4.1
i/t KE, (M- t s- ') ASA
NADP P, V~ E t (s - t )
5.95" 10 f' 1.24" l05 2.32
1260
2.63" I04 1.05" l0 "a 0.47
230 120 490
' Assay conditions: 100 mM Ches ( p t l 8.7) and 0.5 mM D T T and varying concentrations of ASA at several fixed concentrations of N A D P and saturating levels of P, yielded the kinetic parameters for ASA and N A D P from a fit of the data to Eqn. I. The kinetic parameters for Pi were determined from a fit of the data to Eqn. 2 where Pi was varied at saturating levels of A S A and NADP. t, Numbers in p a r e n t h e s e s are the standard errors associated with the kinetic p a r a m e t e r as determined by a fit of the data to Eqn. 1 o r 2.
conditions results in the loss of 98% of the enzyme activity after 15 rain. Kinetic parameters were determined for both the native cysteine enzyme and the Ser-135 mutant. The kinetic parameters for ASA and NADP were derived from an initial velocity pattern run at pH 8.7 by varying ASA at several fixed concentrations of NADP and saturating P~. The initial velocity data were fit to Eqn. 1 to yield the parameters listed in Table II. Initial velocities were determined for P~ at pH 8.7 at varied concentrations of P~ and fixed saturating concentrations of ASA and NADP. These data were fitted to Eqn. 2 to give the parameters presented in Table I! for Pi. The maximum velocity (VIE t) values determined by the !
.r---
[ -
t
6o 5.0
*
4 •
I
4.0
t
ao
i I I I.J
2.0 L
L. . . . 6
t 7
___: 8
9
.......
10
pH
Fig. I. V / K pH profile for the Wild Type (11) and the Ser-135 mutant ( o ) of aspartate-/]-semialdehyde dehydrogenase at saturating levels of N A D P and Pi. T h e data at each p H were fitted to Eqn. 2 and the log of the d e t e r m i n e d kinetic parameters plotted vs. pH. T h e curves through the data are the non-linear fit to Eqn. 3 (Set-135 m u t a n t ) and Eqn. 4 (wild type enzyme).
variation of P~ for the wild type and for the serine mutant enzyme are consistent with the kinetic parameters determined by the variation of ASA and NADP (Table If). To further compare the wild type and the Ser-135 mutant enzymes, the variation of kinetic parameters was examined as a function of pH for these enzymes. A loss of enzyme activity is observed for each enzyme in the acid region of the V/K pH profile with a limiting slope of + 1 (Fig. 1). A fit of the C135S enzyme data to Eqn. 3 yields a pK~ value of 7.7 _+ 0.2 for the enzymatic group that becomes protonated leading to the loss of activity. This pK~ value is shifted up to pH 8.4 + 0.2 in the Vm,, profile (data not shown). For the native enzyme the acid pK a value in the V / K profile is also 7.7 and is about 8.7 in the V,,~x profile. Discussion
it has been proposed that the reaction catalyzed by ASA-DH is initiated by nucleophilic attack on the substrate carbonyl group of ASA by an active site cysteine thiolate anion [5]. Based on the enzyme inactivation results of Biellmann [6] with an alkylating substrate analog of ASA, and the subsequent nucleotide sequence of the asd gene determined by Haziza [1], the most likely candidate for the active site nucleophile has been deduced to be Cys-135. Complete ioss of enzyme activity is observed as a result of the substitution of alanine for this cysteine in ASA-DH. This result clearly indicates an essential role in the enzyme activity for this amino acid residue. While the alanine mutant is inactive the corresponding serine mutant, where Cys-135 is replaced by ~rine, does display some enzymatic activity. "l he maximum velocity of the serine mutant is reduced by about 500-fold compared to the wild type enzyme. Similar decreases are observed in the V/K values, with the magnitude of the decrease depending on which substrate is varied. The KAsA values determined for the serine mutant enzyme at the different pH values in the pH profile are, in most cases, very similar to the wild type Km values at each pH (data not shown). The K m for phosphate is actually. 4-fold lower in the Ser-135 mutant. These results suggest that there are no major conformational changes resulting from the serine substitution that adversely affect the ability of the enzyme to bind its substratcs. The absence of major changes in conformation is also supported by the identical behavior with wild type enzyme displayed by the serine mutant during the purification procedure, in addition, the fluorescence spectra of the wild type and the serine mutant were observed to be identical. These results indicate that the decrease in activity displayed by the serine mutant enzyme is not the result of any significant global structural changes. This is an important point to establish if the kinetic changes observed in the C135S mutant are
238 to be interpreted as being mechanistically relevant. While it would seem that a single conservative amino acid replacement in an enzyme macromolecule should result in only local perturbations, detailed analysis of some enzyme systems has not always confirmed this intuition. The replacement of a single glutamic acid in staphylococcal nuclease with an aspartic acid, as well as with several neutral amino acids, resulted in a dramatic decrease in V/K values and significant, but less drastic increases in K m values [11]. However, when the 2-D NOESY spectra of the wild type and mutant enzymes were compared, differences in the NOE intensities were observed between aromatic protons that are remote from the site of the mutation and adjacent methyl protons. These changes in NOE are indicative of global changes in the conformation of staphylococcal nuclease as a consequence of these single amino acid replacements, and preclude a straightforward interpretation of the role of this active site glutamic acid in catalysis. The lower activity that is observed in the case of the ASA-DH mutant is most likely due to a combination of subtle changes in structure at the active site of the serine mutant and the differences in the reactivity on the enzyme between the cysteine thiol and the serine hydroxyl groups. The chemical mechanism proposed for wild type ASA-DH follows a three step sequence. After attack of the active site thiol, a hydride transfer is postulated from the thiohemiacetal intermediate to NADP. Subsequent attack of an oxygen anion of inorganic phosphate on the resulting thioester intermediate would lead to expulsion of the thiol group and the phosphorylated product. Substitution of a serine hydroxyl group for the native thiol could affect any of these proposed steps to account for the reduction in Vm~, seen with the Ser-135 mutant enzyme. The decrease in Vma~ observed for the Ser-135 mutant enzyme could also be a result of modest changes in the location of the nucleophile as a result of the smaller size of oxygen as compared to sulfur. This could lead to improper positioning of the substrates relative to the active site nucleophile for efficient catalysis to occur. The pK a values in the acid region of the V/K and Vm~, pH profiles are identical for both the native and the ser-135 mutant enzymes. This pK a value is shifted up from about pH 7.7 in the V/K profile to about 8.4 in the Vm~,, profile of the Ser-135 mutant and to about 8.7 in the Vmax profile of the native enzyme [5]. With ASA bound the environment around this enzymatic group apparently becomes more hydrophobic, making ionization of this group more difficult. A shift in a pK a value to higher pH in a less polar environment is the behavior that is expected of a neutral acid group such as a cysteine thiol or a serine hydroxyl group. In solution the pK., for ionization of a cysteine sulhydryl in model peptides is about 4-5 pH units
below the p K,, of a primary alcohol. However, the observed pK~ values are identical for the Set-135 and Cys-135 enzymes. A second amino acid functional group may be located in the active site that deprotonates this nucleophile to set up the enzyme for catalysis. This arrangement of functional groups would be analogous to the thiol-imidazole ion pair present in many cysteine proteinases [12]. A similar ion pair has been proposed for glyceraldehyde-3-phosphate dehydrogenase, whereby His-176 may act as a chemical activator enhancing the reactivity of the active site thiol group [13,14]. The identical PKa values observed at pH 7.7 for the wild type and the serine mutant enzymes in the V/K pH profiles could reflect the p g a value of this putative activating group. Additional studies will be required to examine this aspect of the ASA-DH catalyzed reaction.
Acknowledgements The authors wish to thank Drs. J.J. Villafranca, Spencer Shames and Dr. Jack Preiss for gifts of some of the bacterial strains used in these studies. Dr. Jerry Salem provided technical assistance for the preliminary mutagenesis experiments. We also thank Dr. Paul F. Cook for providing the use of his equipment and supplies for the completion of some of these studies, and Sami Saribas and Jun Ouyang for determining the nucleotide sequences of the genes coding for the mutated enzymes.
References 1 Haziza, C., Stragier, P. and Patte, J.-C. (1982) EMBO J. I, 379-384. 2 Cohen, G.N. (1983) in Amino Acids: Biosynthesis and Genetic Regulation (Herrmann. K.M. and Somerville, R.L., eds.), pp. 147-171, Addi~n-Wesley, Reading. 3 Biellmann, J.-F., Eid, P., Hirth, C. and Jornvall, H. (1980) Eur..I. Biochem. 104, 53-58. 4 Holland, MJ. and Westhead, E.W. (1973) Biochemistry 12, 22642270. 5 Karsten, W.E. and Viola, R.E. (1991) Biochim. 8iophys. Acta, 1077, 209-219. 6 Biellmann, J.-F., Eid, P. and Hirth, C. (1980) Eur. J. Biochem. 104, 59-64. 7 OkRa, T.W., Rodriguez, R.L. and Prciss, J. (1981) J. Biol. Chem. 256, 6944-6952. 8 Preiss, J.. Mazelis, M. and Greenberg, E. (1982) Current Microbiology 7, 263-268. 9 Kunkel, T.A., Roberts, J.D. and Zabour, R.A. (1986) Methods Enzymol. 154E, 367-381. 10 Cleland, W.W. (1967) Adv. Enzymol. 29, 1-32. I I l-lJbler, D.W., Stolowich, N.J., Reynolds, M.A., Gerlt, J.A., Wilde, J.A. and Bolton, P.H. (1987)Biochemistry 26, 6278-6286. 12 Brocklehurst, K. (1987) in Enzyme Mechanisms, pp. 140-158, Royal Society of Chemistry, Burlington House, London. 13 Polgar, L. 0975) Eur. J. Biochem. 52, 63-71. 14 Sourki, A., Mougin, A., Corbier, C., Wonacon, C., Branlant, C. and Branlant, G. (I989) Biochemistry 28, 2586-2592. 15 Bradford, M. (1976) Anal. Biochem. 72, 248-254.
234
BBAPRO 34205
Identification of an essential cysteine in the reaction catalyzed by aspartate-fl-semialdehyde dehydrogenase from Escherichia coil William E. Karsten ' and Ronald E. Viola Departmcn, of Chemist~'. The Unicersit)"of Akron. Akron. OH (USA)
(Received 7 October 1~91) (Revised manuscript received 6 January 1992)
Key words: Enzymemechanism: Oligonucleotide-directed mutagenesis: Enzyme kinetics: Essential cysteine: pH profile The enzyme L-aspartate-/3-semialdehyde dehydrogenase from Eschenchia coil has been studied by oligonucleotide-directed mutagenesis. The focus of this invcstigation was to examine the role of a cysteine residue that had been previously identified by chemical modification with an active sitc directed reagent (Biclimann et al. (1980) Eur. J. Biochem. 104, 59-64). Substitution of this cysteinc at position I35 with an alanine results in complete loss of enzyme activity. However, changing this cysteine to a scrine yields a mutant enzyme with a maximum velocity that is 0.3% that of the native enzyme. This C135S mutant has retained essentially the same affinity for substrates as the native enzyme, and the same overall conformation as reflected in identical behavior on gel electrophoresis and in identical fluorescence spectra. The pH profile of the native enzyme shows a loss in catalytic activity upon protonation of a group with a pK., value of 7.7. The same activity loss is observed at this pH with the ~rine-135 mutant, despite the differences in thc pK a values for a cysteine sulfhydryl and a serine hydroxyl group that have been measured in model compounds. This observed pK~ value may reflect the protonation of an auxmary catalyst that enhances the reactivity of the active site cysteinc nucleophile in the native aspartate-/3-semialdehyde dehydrogenase.
Introduction
L-Aspartate-/3-semialdehyde dehydrogenase (E.C.I. 2.1.11) is a dimeric enzyme composed of identical subunits. The DNA sequence [1] predicts a subunit composed of 367 amino acids with an overall molecular weight of 39950. Starting from L-aspartate, the amino acid biosynthetic pathway in Escherichia coil leads to the formation of L-lysine, L-isoleucine, L-methionine and L-threonine. The first and third steps in this pathway are catalyzed by the bifunctionai enzymes aspartokinase - homoserine dehydrogenase I and II. A third monofunctional enzyme, aspartokinase Ill, also catalyzes the initial reaction in the pathway, and each of these enzymes are differentially regulated to control
the levels of each end product amino acid [2]. L-Aspartate-/~-semialdehyde dehydrogenase (ASA-DH)catalyzes the intervening branch point reaction between the aspartokinase and the homoserine dehydrogenase reactions, with one of these branches leading to the production of L-lysine and the other leading to Lmethionine, L-threonine and 1.-isoleucine. In the biosynthetic direction ASA-DH catalyzes the formation of L-aspartate-/3-semialdehyde (ASA) by the reduetive dephosphorylation of L-/3-aspartyl-phosphate (BAP) utilizing NADPH: COO ~ H3N_CH+NADPH+H + i
C(O)OPOi i Present address: Department of Microbiology and Immunology, Texas College of Osteopathic Medicine, Fort Worth, TX, USA. Abbreviations: ASA, aspartate-~-semialdehyde: ASA-DH, asparlate/3-semialdehyde dehydrogenase, BAP, /3-aspartylphosphate: DTNB, 5,5'-dithiobis42-nitrobenzoic acid): DTI', dithiothreilol; NEM, Nethylmaleimide. Correspondence: R.E. Viola, Department of Chemistry, The University of Akron, Akron, OH 44325-3601, USA.
,43-asparlyl phosphate
COO
-
÷H~N_CH+H2PO4-+NADP ÷ "
~{C-O i -asparlatc-/3. -scmialdehyde
The ASA-DH reaction is related to the reaction catalyzed by glyceraldehyde-3-phosphate dehydrogenase and similar chemical mechanisms have been proposed for each enzyme [3-5]. The suggested mechanism involves the formation of a thioester intermediate resulting from attack of a cysteine thiolate on the
235 substrate carbonyl group, followed by hydride transfer to NADP. Subsequent attack on the thioester intermediate by an oxygen anion of bound inorganic phosphate leads to expulsion of the cysteine thiolate and formation of the phosphorylated product. Chemical modification studies with DTNB [31 and N-ethylmaleimide [5] have indicated an essential role for a cysteine thiol in the ASA-DH reaction_.,. Substrate protection studies against N-ethylmaleimide inactivation performed with ASA have suggested that the essential cysteine is lo-cated in or near the enzyme active site. In addition, pH studies have indicated a role for a neutral acid group that must be ionized for enzymatic activity. The p K a value determined for this group is similar t-q that for the enzymatic group that leads to loss of enzyme activity when modified by N-ethylmaleimide [5]. L-2-Amino4-oxo-5-chloropentanoic acid, an alkylating substrate analogue of ASA, has been shown to irreversibly inactivate ASA-DH. An isolated peptide containing the modified amino acid, tentatively identified as a histidine, was reported to have the amino acid sequence Phe-Val-Gly-Gly-Asp-His-Thr-Val-Ser [6]. Haziza [1] determined the nucleotide sequence of the asd gene of E. coli that codes for ASA-DH and found a corresponding sequence encoding for amino acid residues 130-138 of their deduced sequence. However, they found a substitution of asparagine for aspartic acid and cysteine for histidine at residues 134 and 135. These results are consistent with the cysteine located at position 135 of the amino acid sequence being the most likely candidate for the active site cysteine. Utilizing site-directed mutagenesis we have constructed two mutants of ASA-DH in order to further clarify the role of cysteine 135 in the chemical mechanism of the enzyme. We report here on the preparation, activity and characterization of these mutant enzymes, and on the implications of these results on the chemical mechanism of ASA-DH.
Experimental procedures Materials. Restrict;on enzymes were purchased from Promega or U.S. Biochemical Corporation. T4 DNA ligase, MI3K07 bacteriophage, bacterial strain JM109 and the pGEM-7Zf( + ) plasmid vector were purchased from Promega. The strain of E. coil harboring the plasmid pOP126 containing the asd gene [7,8] was a gift from Dr. Spencer Shames. The E. coli strains RZ1032 and HBI01 were obtained from Dr. J.J. Villafranca. The deletion mutant devoid of ASA-DH activity, E. coli KI2 G6MD3 [7], was a gift from Dr. Jack Preiss. Unpurified oligonucleotides were purchased from Operon Technologies. Sequencing of DNA was performed by the dideoxynueleotide chain termination method using the Sequenase DNA sequencing kit purchased from U.S. Biochemical Corporation.
Bacterial growth and media. A standard enriched media was used for growth of all bacteria and contained Bacto-tryptone (8 g/l), yeast extract (5 g/l) and sodium chloride (2.5 g/l). E. coli strain K12 G6MD3 requires diaminopimelate, and was grown on media containing 50 # g / m l diaminopimelate. Ampicillin was included at a concentration of 100 # g / m l where required. Subcloning and mutagenesis. The plasmid pOP126 contains the asd gene on an approx. 1600 base pair fragment inserted into the Bam H I site of pBR322. This BamH ! fragment from pOP126 was subcloned into the Bam H ! site of the pGEM-7Zf(+ ) plasmid vector to conduct the mutagenesis studies. The Kunkel site-directed mutagenesis method [9] utilizing uracilcontaining ssDNA was performed on the pGEM subclone of the asd gene. The cysteine-135 codon (TGT) was replaced with either alanine (GCT) or serine (TCG) by using 5'-GCTI'ACGGTAGCGTTACCGCC-3' as the primer for the alanine mutation (base changes underlined), and the corresponding 21-base oligonucleotide for the serine mutation. The substitution of each of these two nucleotides for the respective wild type bases also eliminates an Rsal restriction enzyme site that overlaps the TGT cysteine codon in the native DNA. Rsal restriction enzyme digests to detect the presence or absence of this restriction enzyme site were used to carry out the initial screening of plasmids for the presence of mutant DNA. The amino acid substitutions were verified by determination of the nucleotide sequence in the region of the gene that codes for position-135. Sequencing was initiated within several hundred bases of this position by utilizing synthetic oligonueleotides as primers. Characterization of enzymes. The C135S and wild type enzymes were purified by the method of Karsten and Viola [5]. Protein concentrations were determined by the method of Bradford [15]. Enzyme activity was routinely assayed in the phosphorylating direction at 25°C in 1 ml cuvettes containing varying concentrations of substrates and 0.5 mM DTT. For the variation of kinetic parameters as a function of pH the buffers used were Mes (pH 6.0-6.5), Hepes (pH 7.0-8.0), and Ches or Taps (pH 8.5-9.5) at a concentration of 100 mM. Enzyme assays were performed on a Gilford 260 recording spectrophotometer equipped with thermospacers and connected to a circulating water bath to maintain a constant temperature. The fluorescence spectra of the native and thc mutant enzymes were examined on a Shimadzu RF-5000U fluorescence spectrometer in 50 mM Hepes (pH 7.0), with 0.5 mM DTT. The enzyme solutions were excited at 280 nm and the emission spectra were recorded from 300 to 400 nm. Chemical modification studies with N-ethylmaleimide were carried out in 100 mM Hepes (pH 7.0). From 7-20 ~g of enzyme (either
236 native or C135S mutant) was incubated with 0.2 mM NEM. Aliquots were removed at various times and examined for enzymatic activity in an assay mixture in 150 mM Ches (pH 9.0) containing 0.1 mM ASA, 0.2 mM NADP, 25 mM potassium phosphate and 0.5 mM DTI'. Data analysis. The kinetic data were analyzed by using BASIC versions of the kinetics programs of Cleland [10]. Initial velocity data giving intersecting patterns were fitted to Eqn. 1 where c = initial velocity; V= maximum velocity; A or B = substrate concentrations; K = Michaelis constants. c=
VAB KiaK~ + Ka B + KbA + AB
(I)
For the pH profiles kinetic parameters were determined at each pH by varying the concentration of ASA at saturating levels of the other two substrates and fitting the initial velocity data to Eqn. 2. For the serine mutant, the V / K values determined at each pH were fitted to Eqn. 3, which allows for activity decreasing in the acid region of the pH profile, and where C is the maximum V / K value and K~ is the acid dissociation constant. For the native enzyme, the V / K values were fitted to Eqn. 4, which allows for a decrease in activity in both the acid and basic regions. VA
c=
K+A
(2)
log V / K = log{C/( I + [H " I/K..,)}
(3)
log V / K = log{C,/( I -)-[H" I/K., + K h / I H ÷ l)}
(4)
Results
Construction and purification of mutants Mutants of ASA-DH were constructed by oligonucleotide-directed mutagenesis to examine the role of a putative active site cysteine in the mechanism of the enzyme-catalyzed reaction. Initial screening for the presence of the Ala-135 or Ser-135 mutant was accomplished by Rsal restriction enzyme digests of isolated plasmids. The loss of the Rsal restriction enzyme site that overlaps the wild type cysteine codon was taken as evidence for the presence of the desired mutation. Plasmids producing the appropriate restriction map were then sequenced through the region coding for the amino acid at position 135, by using the dideoxy chain termination method, to verify the identity of the amino acid produced at that position. Plasmids that were identified to contain the desired mutation were used to transform the deletion mutant G6MD3 E. coli strain that is essentially devoid of ASA-DH activity. Cultures of these transformed cells were grown overnight, and then disrupted by sonic oscillation. The crude extracts from these transformed cells were assayed for the
TABLE I Acticity of ASA-DH mutants
Enzyme"
Wild type C 135A C135S
Activity h pH 7
(%)
pH 9
(%)
2.1 0.01)06 (I.0049
(100) (0.(13) (0.2)
15.7 0.00016 0.052
(100) (0.001) (0.3)
Extracted from the G6MD3 strain of E. coli containing the pGEM plasmid with an insert of either the wild type or a mutated asd gene. h Assay conditions: 200 mM Hepes (pH 7.0) or 200 mM Ches (pH 9.0), 0.2-[I.5 mM NADP, 0.15 mM ASA, 50 mM Pi and 0.5 mM DTI'. Enzyme activity in crude cell extracts, expressed in 0,tool/rain per mg of total protein, or as a percentage of the wild type activity at that pH.
acquisition of ASA-DH activity. The Ala-135 mutant displays some very low ASA-DH activity in crude extracts at pH 7, however, the level of activity is virtually identical to that observed in the G6MD3 strain that does not contain the plasmid. Essentially no enzyme activity is observed in the Ala-135 mutant strain at pH 9 (Table !), which is the pH optimum for wild type enzyme activity. Also, partial purification of this C135A mutant protein resulted in the loss of this background activity. The Ser-135 mutant enzyme showed activity at both pH 7 and 9, but the activity that was measured was well below the wild type levels that is typically observed in these crude extracts. Samples of the crude extracts obtained from each of these mutant strains were run on SDS-PAGE and stained with Coumassie blue. In each case a protein band comprising approx. 15% of the total protein was present in the sample at a position identical with that of the purified wild type ASA-DH. The Ser-135 mutant enzyme was purified from these crude extracts in order to more fully characterize the activity of this enzyme. The mutant enzyme behaved in an identical fashion to the wild type enzyme throughout the purification. The purified Ser-135 mutant was determined to be > 95% pure by Coumassie blue stained protein gels.
Characterization of the Ser-135 mutant enzyme The fluorescence spectra of the native and the CI35S enzymes were examined from 300 to 400 nm. The maxima for each enzyme was observed at 320 rim, and the spectra of each were superimposeable (data not shown). Reaction of N-ethylmaleimide (NEM)with the native enzyme results in a rapid and complete loss of enzyme activity with a corresponding modification of a single cysteine residue [5]. When the serine-135 mutant is treated with NEM there is only about a 15% decrease in activity after incubation with the reagent for 15 rain. Treatment of the native enzyme under these
237 T A B L E II
Comparison of the kinetic parameters between wild type and Ser-135 mutant of ASA-DH Parameter
Wild type
K m (mM) ~ ASA NADP P+
C135S
Ratio w.t./mutant
0.057 (0.{172) b 0.038 (0.017) 1.86 (0.33)
0.146 (0.02b) 0.018 (0.005) 0.45 (0.1)
4.07. l i p
3.24-103
0.4 2.1
4.1
i/t KE, (M- t s- ') ASA
NADP P, V~ E t (s - t )
5.95" 10 f' 1.24" l05 2.32
1260
2.63" I04 1.05" l0 "a 0.47
230 120 490
' Assay conditions: 100 mM Ches ( p t l 8.7) and 0.5 mM D T T and varying concentrations of ASA at several fixed concentrations of N A D P and saturating levels of P, yielded the kinetic parameters for ASA and N A D P from a fit of the data to Eqn. I. The kinetic parameters for Pi were determined from a fit of the data to Eqn. 2 where Pi was varied at saturating levels of A S A and NADP. t, Numbers in p a r e n t h e s e s are the standard errors associated with the kinetic p a r a m e t e r as determined by a fit of the data to Eqn. 1 o r 2.
conditions results in the loss of 98% of the enzyme activity after 15 rain. Kinetic parameters were determined for both the native cysteine enzyme and the Ser-135 mutant. The kinetic parameters for ASA and NADP were derived from an initial velocity pattern run at pH 8.7 by varying ASA at several fixed concentrations of NADP and saturating P~. The initial velocity data were fit to Eqn. 1 to yield the parameters listed in Table II. Initial velocities were determined for P~ at pH 8.7 at varied concentrations of P~ and fixed saturating concentrations of ASA and NADP. These data were fitted to Eqn. 2 to give the parameters presented in Table I! for Pi. The maximum velocity (VIE t) values determined by the !
.r---
[ -
t
6o 5.0
*
4 •
I
4.0
t
ao
i I I I.J
2.0 L
L. . . . 6
t 7
___: 8
9
.......
10
pH
Fig. I. V / K pH profile for the Wild Type (11) and the Ser-135 mutant ( o ) of aspartate-/]-semialdehyde dehydrogenase at saturating levels of N A D P and Pi. T h e data at each p H were fitted to Eqn. 2 and the log of the d e t e r m i n e d kinetic parameters plotted vs. pH. T h e curves through the data are the non-linear fit to Eqn. 3 (Set-135 m u t a n t ) and Eqn. 4 (wild type enzyme).
variation of P~ for the wild type and for the serine mutant enzyme are consistent with the kinetic parameters determined by the variation of ASA and NADP (Table If). To further compare the wild type and the Ser-135 mutant enzymes, the variation of kinetic parameters was examined as a function of pH for these enzymes. A loss of enzyme activity is observed for each enzyme in the acid region of the V/K pH profile with a limiting slope of + 1 (Fig. 1). A fit of the C135S enzyme data to Eqn. 3 yields a pK~ value of 7.7 _+ 0.2 for the enzymatic group that becomes protonated leading to the loss of activity. This pK~ value is shifted up to pH 8.4 + 0.2 in the Vm,, profile (data not shown). For the native enzyme the acid pK a value in the V / K profile is also 7.7 and is about 8.7 in the V,,~x profile. Discussion
it has been proposed that the reaction catalyzed by ASA-DH is initiated by nucleophilic attack on the substrate carbonyl group of ASA by an active site cysteine thiolate anion [5]. Based on the enzyme inactivation results of Biellmann [6] with an alkylating substrate analog of ASA, and the subsequent nucleotide sequence of the asd gene determined by Haziza [1], the most likely candidate for the active site nucleophile has been deduced to be Cys-135. Complete ioss of enzyme activity is observed as a result of the substitution of alanine for this cysteine in ASA-DH. This result clearly indicates an essential role in the enzyme activity for this amino acid residue. While the alanine mutant is inactive the corresponding serine mutant, where Cys-135 is replaced by ~rine, does display some enzymatic activity. "l he maximum velocity of the serine mutant is reduced by about 500-fold compared to the wild type enzyme. Similar decreases are observed in the V/K values, with the magnitude of the decrease depending on which substrate is varied. The KAsA values determined for the serine mutant enzyme at the different pH values in the pH profile are, in most cases, very similar to the wild type Km values at each pH (data not shown). The K m for phosphate is actually. 4-fold lower in the Ser-135 mutant. These results suggest that there are no major conformational changes resulting from the serine substitution that adversely affect the ability of the enzyme to bind its substratcs. The absence of major changes in conformation is also supported by the identical behavior with wild type enzyme displayed by the serine mutant during the purification procedure, in addition, the fluorescence spectra of the wild type and the serine mutant were observed to be identical. These results indicate that the decrease in activity displayed by the serine mutant enzyme is not the result of any significant global structural changes. This is an important point to establish if the kinetic changes observed in the C135S mutant are
238 to be interpreted as being mechanistically relevant. While it would seem that a single conservative amino acid replacement in an enzyme macromolecule should result in only local perturbations, detailed analysis of some enzyme systems has not always confirmed this intuition. The replacement of a single glutamic acid in staphylococcal nuclease with an aspartic acid, as well as with several neutral amino acids, resulted in a dramatic decrease in V/K values and significant, but less drastic increases in K m values [11]. However, when the 2-D NOESY spectra of the wild type and mutant enzymes were compared, differences in the NOE intensities were observed between aromatic protons that are remote from the site of the mutation and adjacent methyl protons. These changes in NOE are indicative of global changes in the conformation of staphylococcal nuclease as a consequence of these single amino acid replacements, and preclude a straightforward interpretation of the role of this active site glutamic acid in catalysis. The lower activity that is observed in the case of the ASA-DH mutant is most likely due to a combination of subtle changes in structure at the active site of the serine mutant and the differences in the reactivity on the enzyme between the cysteine thiol and the serine hydroxyl groups. The chemical mechanism proposed for wild type ASA-DH follows a three step sequence. After attack of the active site thiol, a hydride transfer is postulated from the thiohemiacetal intermediate to NADP. Subsequent attack of an oxygen anion of inorganic phosphate on the resulting thioester intermediate would lead to expulsion of the thiol group and the phosphorylated product. Substitution of a serine hydroxyl group for the native thiol could affect any of these proposed steps to account for the reduction in Vm~, seen with the Ser-135 mutant enzyme. The decrease in Vma~ observed for the Ser-135 mutant enzyme could also be a result of modest changes in the location of the nucleophile as a result of the smaller size of oxygen as compared to sulfur. This could lead to improper positioning of the substrates relative to the active site nucleophile for efficient catalysis to occur. The pK a values in the acid region of the V/K and Vm~, pH profiles are identical for both the native and the ser-135 mutant enzymes. This pK a value is shifted up from about pH 7.7 in the V/K profile to about 8.4 in the Vm~,, profile of the Ser-135 mutant and to about 8.7 in the Vmax profile of the native enzyme [5]. With ASA bound the environment around this enzymatic group apparently becomes more hydrophobic, making ionization of this group more difficult. A shift in a pK a value to higher pH in a less polar environment is the behavior that is expected of a neutral acid group such as a cysteine thiol or a serine hydroxyl group. In solution the pK., for ionization of a cysteine sulhydryl in model peptides is about 4-5 pH units
below the p K,, of a primary alcohol. However, the observed pK~ values are identical for the Set-135 and Cys-135 enzymes. A second amino acid functional group may be located in the active site that deprotonates this nucleophile to set up the enzyme for catalysis. This arrangement of functional groups would be analogous to the thiol-imidazole ion pair present in many cysteine proteinases [12]. A similar ion pair has been proposed for glyceraldehyde-3-phosphate dehydrogenase, whereby His-176 may act as a chemical activator enhancing the reactivity of the active site thiol group [13,14]. The identical PKa values observed at pH 7.7 for the wild type and the serine mutant enzymes in the V/K pH profiles could reflect the p g a value of this putative activating group. Additional studies will be required to examine this aspect of the ASA-DH catalyzed reaction.
Acknowledgements The authors wish to thank Drs. J.J. Villafranca, Spencer Shames and Dr. Jack Preiss for gifts of some of the bacterial strains used in these studies. Dr. Jerry Salem provided technical assistance for the preliminary mutagenesis experiments. We also thank Dr. Paul F. Cook for providing the use of his equipment and supplies for the completion of some of these studies, and Sami Saribas and Jun Ouyang for determining the nucleotide sequences of the genes coding for the mutated enzymes.
References 1 Haziza, C., Stragier, P. and Patte, J.-C. (1982) EMBO J. I, 379-384. 2 Cohen, G.N. (1983) in Amino Acids: Biosynthesis and Genetic Regulation (Herrmann. K.M. and Somerville, R.L., eds.), pp. 147-171, Addi~n-Wesley, Reading. 3 Biellmann, J.-F., Eid, P., Hirth, C. and Jornvall, H. (1980) Eur..I. Biochem. 104, 53-58. 4 Holland, MJ. and Westhead, E.W. (1973) Biochemistry 12, 22642270. 5 Karsten, W.E. and Viola, R.E. (1991) Biochim. 8iophys. Acta, 1077, 209-219. 6 Biellmann, J.-F., Eid, P. and Hirth, C. (1980) Eur. J. Biochem. 104, 59-64. 7 OkRa, T.W., Rodriguez, R.L. and Prciss, J. (1981) J. Biol. Chem. 256, 6944-6952. 8 Preiss, J.. Mazelis, M. and Greenberg, E. (1982) Current Microbiology 7, 263-268. 9 Kunkel, T.A., Roberts, J.D. and Zabour, R.A. (1986) Methods Enzymol. 154E, 367-381. 10 Cleland, W.W. (1967) Adv. Enzymol. 29, 1-32. I I l-lJbler, D.W., Stolowich, N.J., Reynolds, M.A., Gerlt, J.A., Wilde, J.A. and Bolton, P.H. (1987)Biochemistry 26, 6278-6286. 12 Brocklehurst, K. (1987) in Enzyme Mechanisms, pp. 140-158, Royal Society of Chemistry, Burlington House, London. 13 Polgar, L. 0975) Eur. J. Biochem. 52, 63-71. 14 Sourki, A., Mougin, A., Corbier, C., Wonacon, C., Branlant, C. and Branlant, G. (I989) Biochemistry 28, 2586-2592. 15 Bradford, M. (1976) Anal. Biochem. 72, 248-254.
