- Email: [email protected]
The binding of propionyl-CoA and carboxymethyl-CoA to Escherichia coli citrate synthase
ELSEVIER
Biochi~ic~a et Biophysica ~ t a Biochimicaet BiophysicaActa 1250 (1995) 69-75
The binding of propionyl-CoA and carboxymethyl-CoA to Escherichia coli citrate synthase Wai-Jin Man, Yan Li, C. David O'Connor, David C. Wilton * Departmentof Biochemistry, Universityof Southampton,BassettCrescentEast, SouthamptonS09 3TU, UK Received 4 November1994; revised 13 February 1995; accepted7 March 1995
Abstract The interaction of propionyl-CoA and acetyl-CoA with E. coli citrate synthase has been studied in order to gain insight into the structural requirements for sabstrate binding by this enzyme. In contrast to the enzyme from pig heart, the E. coli enzyme was unable to catalyse significant exchange of the methylene protons of propionyl-CoA while overall activity was very low with this enzyme. Carboxymethyl-CoA is a presumptive transition state analogue of acetyl-CoA using pig heart citrate synthase. The effect of carboxymethyl-CoA on both the native enzyme from E. coli and a catalytically active aspartate mutant (D362E) was investigated. Whereas the native enzyme was inhibited by carboxymethyl-CoA, the mutant enzyme (D362E) shows either no inhibition or minimal inhibition depending on the assay conditions. The binding of acetyl-CoA is not inhibited as a result of the mutation. The results with propionyl-CoA and carboxymethyl-CoA suggest that the active site of the E. coli enzyme is more restricted as compared with the enzyme from pig heart and, in the case of propionyl-CoA, this restriction prevents the formation of a catalytically productive enzyme-substrate complex.
Keywords: Citrate synthase; Propionyl-CoA;Carboxymethyl-CoA;Mutagenesis
1. Introduction
Citrate synthase belongs to an important group of enzymes that catalyse the formation of a new carbon-carbon bond. The enzyme catalyses the following Claisen condensation reaction: acetyl - CoA + oxaloace~:ate + H 2 0 ~ citrate + CoA + H ÷ The reaction can be divided into at least 2 steps, namely the condensation reaction to produce enzyme bound citrylCoA and the hydrolysis of citryl-CoA to produce products. Deprotonation of acetyl-CoA is a crucial event in the initial condensation reaction. Oxaloacetate binds first causing a conformational change which results in the formation of a well defined binding site for acetyl-CoA allowing subsequent attack on the si-face of oxaloacetate to form S-citryl-CoA. Most studies on citrate synthase have been carried out with the pig heart enzyme for which the crystal structure
* Corresponding author. Fax: +44 703 594459. 0167-4838/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0167-4 83 8(95)0004,~.-5
was determined by Huber's laboratory [1]. However, the identified active site residues in this enzyme are conserved in all species of citrate synthase so far studied indicating that a common mechanism is involved. A crystal structure of the enzyme from pig heart involving a ternary complex with oxaloacetate and carboxymethyl-CoA, a presumptive transition state analogue of the enol form of acetyl-CoA, allowed the proposal of an attractive mechanism for the condensation reaction [2] involving specific active-site residues. In this mechanism the conserved active-site aspartate (Asp-375) acts as a base to deprotonate the methyl carbon acetyl-CoA. This process is facilitated by the active site histidine (His-274) protonating the carbonyl oxygen to produce the enol form of acetyl-CoA [2]. This proposal was supported by mutagenesis studies [3] and further crystallographic studies involving ternary complexes with D- or L-malate and acetyl-CoA [4]. Recently, additional support for the mechanism has come from crystallographic studies using the acetyl-CoA analogues, carboxymethyldethia CoA and amidocarboxymethyldethia CoA [5].
70
W.-J. Man et al. / Biochimica et Biophysica Acta 1250 (1995) 69-75
It is apparent that much of the experimental basis for the proposed mechanism has been derived from the crystal structures of CoA analogues bound to the pig heart enzyme. Although a crystal structure of the E. coli enzyme is not available, mutagenesis studies by us involving the conserved active site aspartate suggested a different mechanism [6]. This mechanism did not involve the catalytic aspartate in the condensation stage of the reaction. Because the mechanism proposed for the pig heart enzyme was based initially on crystallographic studies using carboxymethyl-CoA, we have investigated the binding of this acetyl-CoA analogue to both wild-type and the aspartate mutant (D362E) of E. coli citrate synthase. In addition, another acetyl-CoA analogue, propionyl-CoA, can be metabolised by the pig heart enzyme and this is accompanied by rapid exchange of the ce-hydrogens of this substrate [7]. This exchange reaction should provide a method for measuring substrate deprotonation separate from overall catalysis in the E. coli enzyme. It is the mechanism of this deprotonation step that requires clarification using the E. coli enzyme. Overall, the results using the E. coli enzyme indicated that the binding site for acetyl-CoA with this citrate synthase is more catalytically restricted than that found in the pig heart enzyme. In particular, no detectable proton exchange using propionyl-CoA could be detected using the enzyme from E. coli and therefore it was not possible to investigate the role of the active site aspartate in the process. In view of these differences between the two enzymes, it must remain a possibility that mechanisms proposed for the E. coli enzyme may not be applicable to eukaryotic citrate synthases.
2. Materials and methods
2.1. Materials Carboxymethyl-CoA was prepared as described by Bayer et al. [8]. All other materials were obtained or prepared as described previously [6].
2.2. Site-directed mutagenesis The 2.1 kb SacI-SalI gltA gene encoding E. coli citrate synthase from plasmid pDB2 [9] was cloned into M13mpl9. Mutagenesis was carried according to the Kunkel method [10] using the mutagenic primer, mpl (D362E): 5'-CCGAACGTCGAATTCTACTCT-3' which was purified as previously described [11]. The bacteriophage containing the mutated gltA gene was re-sequenced with the use of specific primers before recloning the gene, as an SacI-BgllI fragment, back into pDB2. E. coli strain DEKI5, which is recombination deficient and carries a deletion of the gltA gene, was employed to harvest the mutated pDB2 plasmid.
2.3. Purification of wild-type and mutant citrate synthases Expression and purification of these enzymes was carried out as described previously [6].
2.4. Enzyme assays The 5,5'-dithio-bis(2-nitrobenzoate) (DTNB) assay of citrate synthase and NADH binding assays have been described previously [6,12]. The E. coli enzyme demonstrates allosteric properties and may be activated from Tto R-form by 0.1 M KC1 which enhances acetyl-CoA and NADH binding [13,14]. The effect of salt concentration and NADH binding of the active-site D362E mutant, has been investigated [15] and the results suggest a small shift in the equilibrium of the D362E mutant in favour of the R-state. Therefore, assays involving the binding of carboxymethyl-CoA and NADH were performed in both the absence and presence of 0.1 M KC1. All spectrophotometric assays were performed at room temperature in a total volume of 0.8 ml using an Hitachi U2000 spectrophotometer. NADH binding studies were performed using an Hitachi F2000 fluorimeter. •
3
2.5. Preparanon of H-labelled propzonyl-CoA with citrate synthase and oxaloacetate Propionyl-CoA (5 /zmol) was incubated for 40 min in 3H20 (100 mCi) in the presence of oxaloacetate (5 ~mol) and pig heart citrate synthase (0.1 mg). A total volume of 1 ml was used. The reaction mixture was freeze-dried, the residue was dissolved in dilute HCI and freeze-dried again. The residue was redissolved in 3 mM HC1 and loaded onto a DEAE-Sephacel column (1 × 25 cm), equilibrated with 3 mM HCI. Propionyl-CoA was eluted with a linear gradient formed between 150 ml of 20 mM LiC1 in 3 mM HC1 and 150 ml of 150 mM LiC1 in 3 mM HC1. 2 ml fractions were collected and monitored at 257 nm for the presence of the coenzyme A moiety and tritiated material was detected using scintillation counting. The peak corresponding to [3H]propionyl-CoA was isolated and freeze-dried. The specific activity of the propionyl-CoA was 5 . 1 . 105 dpm//zmol.
2.6. [3H]Propionyl-CoA as a substrate for wild-type and mutant E. coli citrate synthase 10-20 /zg of wild-type or D362E mutant citrate synthase was incubation in the presence of 1 mM propionylCoA, 30000 dpm of [3H]propionyl-CoA, 1 mM oxaloacetate, 100 /xM DTNB and Tris-HCl buffer (pH 8) for 1-120 min. The reactions were monitored spectrophotometrically at 412 nm and terminated using 0.05 ml of 1 M KOH and 0.01 ml of 1 M sodium acetate. The reaction mixtures were lyophilised and the radioactivity in 0.2 ml of the 3H-labelled water was determined.
71
W.-J. Man et aL / Biochimica et Biophysica Acta 1250 (1995) 69-75 100
2.7. [ 3H]Propionyl-CoA as an inhibitor for wild-type and mutant E. coli citrate synthase
The inhibitory effect of propionyl-CoA on wild-type and the D362E mutant citrate synthase was studied. The normal DTNB assay was used with the addition of 500 /zM propionyl-CoA. The J~'i for propionyl-CoA was determined by varying the concentration of oxaloacetate (0-200 #M). The reaction was followed spectrophotometrically at 412 nm using the Hitachi U2000 spectrophotometer.
~
8o
~,
60
E ~
40
2() A 0
3. Results
i
i
i
p
20
40
60
80
100
[ C a r b o x y m e t h y I - C o A ] (~M)
3.1. Propionyl-CoA as a substrate for citrate synthase from pig heart and E. coLi
150 125
In the presence of oxaloacetate pig heart citrate synthase is able to bind propionyl-CoA with high affinity and is able to catalyse the overall reaction to methyl citrate at a rate of about 0.01% of that observed for citrate formation [7]. In addition, the enzyme is able to rapidly exchange the c~-protons of propionyl-CoA with the pro-S hydrogen being exchanged 15 times faster than the pro-R hydrogen. This facile exchange catalysed by the enzyme in the presence of oxaloacetate provided a potential system for investigating the role of the active site aspartate in substrate deprotonation in the case of the E. coli enzyme. Therefore, [2-3H]propionyl-CoA was prepared by making use of the exchange reaction catalysed by pig heart enzyme and the ability of the E. coli enzyme to catalyse this exchange and the overall reaction was investigated. The results of studies using propionyI-CoA and both the wild-type and D362E mutant citrate synthase are shown in Table 1. Under the assay conditions there was a very low but detectable overall activity suggesting formation of methyl citrate in the wild-type enzyme. However, no such activity could be detected in the D362E mutant consistent with the observation that this mutant has only about 1% of the overall activity of the wild-type enzyme [6]. The propionyl-CoA activity s~een with the wild-type enzyme was proportionally similar to that observed with the pig heart enzyme. However, when the ability of the wild-type
~
A
100
~
75
E
50
N C 1.1.1
25 0 20
40
60
80
100
[ C a r b o x y r n e t h y I - C o A ] (pM}
Fig. 1. The effect of carboxymethyl-CoA on the activity of wild-type and mutant citrate synthases. The reactions were performed in the presence of 125/zM acetyl-CoA, 125 /.tM oxaloacetate and 125/zM DTNB in 0.1 M Tris-HC1 buffer (pH 8.0). Assays were performed in the presence (A) and absence (B) of 0.1 M KC1 and activities are expressed as a percentage of control activity in the absence of carboxymethyl-CoA. (-O-) Wild-type enzyme (4.4 p.g), and ( - • - ) D362E mutant enzyme (116 /zg).
E. coli enzyme to release tritium from [2-3H]propionylCoA was studied no significant exchange of tritium could be detected even with large amounts of enzyme. It should be emphasised that the compulsory loss of tritium associated with the very low rate of propionyl-CoA metabolism would not be detected in these experiments using the E. coli enzyme. Under identical conditions there was a considerable release of tritium by the pig heart enzyme. The
Table 1 Propionyl-CoA as the substrate or inhibitor for E. coli citrate synthase Citrate synthase
Activity (U/mg)
Tritium exchange ( / x m o l / m i n per mg)
Propionyl CoA K i (/zM)
Pig heart Wild-type E. coli D362E E. coli
0.002 + 0.0005 < 0.0003 -
3.6 + 0.3 < 0.025 < 0.004
540 + 200 670 + 170
All assays were performed in 0.1 M Tris-HCl buffer at pH 8.0. Assays for enzyme activity with propionyl-CoA as substrate were performed using 250 poM propionyl-CoA and 500 p.M oxaloacetate. Inhibition studies were performed using 250 /.LM acetyl-CoA, 500 /.LM and the concentration of oxaloacetate was varied from 0 - 2 0 0 / x M . 4 ~ g of wild-type and 116/zg D362E mutant E. coli citrate synthase was used. Tritium exchange studies were performed as described in Section 2.
72
W.-J. Man et al. / Biochimica et Biophysica Acta 1250 (1995) 69-75
rate of tritium release for the pig heart enzyme (3.6 /zmol/min per mg) compares favourably with the value of 6.6 ~ m o l / m i n per mg previously reported for this enzyme [7]. In view of the inability of the E. coli enzyme to use propionyl-CoA as a substrate it was of interest to assess the binding of the propionyl-CoA to the enzyme. Since the actual rate observed with the wild-type enzyme was only just within the limit of sensitivity of the normal DTNB-linked assay for this enzyme, binding was assessed in terms of the ability of the propionyl-CoA to act as a competitive inhibitor. The K i values for wild-type and D362E mutant E. coli citrate synthase were 540 /~M and 667 /zM, respectively, indicating that it does bind weakly to the E. coli enzyme. The corresponding K m value for acetyl-CoA under these conditions was 350 /xM [6]. Therefore, the lack of activity exhibited with the E. coli enzyme cannot be explained by the inability of propionyl-CoA to bind to the enzyme.
3.2. The effect of carboxymethyl-CoA on the activity of E. coli citrate synthase The effect of varying concentrations of carboxymethylCoA on the overall activity of citrate synthase wild-type and D362E mutant enzymes was measured under standard conditions in the absence or presence of 0.1 M KCI. The presence of 0.1 M KC1 favours the R-form of the enzyme with an increased affinity for acetyl-CoA [13,14] (see Section 2). In the wild-type enzyme the inhibitory effect of carboxymethyl-CoA is more apparent in the presence of KC1 (Fig. 1A) as compared with the absence of KC1 (Fig. 1B) suggesting that carboxymethyl-CoA binds more avidly to the R-form of the enzyme as is known for acetyl-CoA binding [13]. However, a dramatic effect is seen when the action of carboxymethyl-CoA on the mutant enzyme is
300
A 250
200
150 ,11
100
50
0 0
r
i
i
5
10
15
20
AcetyI-CoA Concentration (xl0 -~ pM-1l 200
B
1so
~
~
100
s0
0 0.00
~
i
i
i
0.01
0.02
0.03
0.04
Oxaloacetate Concentration (pM-1) Fig. 3. Double reciprocal plots of enzyme activity in the absence or presence of carboxymethyl-CoA for wild-type citrate synthase. Assays were performed (A) in the presence of 250 /xM oxaloacetate or (B) 250 /-~M acetyl-CoA using wild-type citrate synthase (4 ixg). All assays were performed in the presence of O. 1 M KCI. (- • - ) No inhibitor, and (- • -) 25 /xM carboxymethyl-CoA.
200
~ w
15o
I00
E N =
50-
J
i
r
,
i
100
200
300
400
500
600
[ C a r b o x y m e t h y I - C o A ] (~M) Fig. 2. The effect of high concentrations of carboxymethyl-CoA on the activity of the D362E mutant citrate synthase. The assay conditions were as described in Fig. 1 except that the concentration of carboxymethyl-CoA used was up to 600 ~ M . (- • -) Assay in the absence of KC1, and (- • -) assay in the presence of 0.1 M KCI.
examined. In the absence of KC1, carboxymethyl-CoA partially activates the enzyme, a result most readily explained by the inhibitor binding at a second site and stabilising the R-form. This conclusion is supported by assaying the effect of carboxymethyl-CoA in the presence of KC1 which will also stabilise the R-form. Under these conditions no activation is observed while significant inhibition can be detected in the presence of high concentrations of carboxymethyl-CoA (Fig. 2). It was not possible to analyze the effects of carboxymethyl-CoA on the activity of the mutant enzyme in the absence of KC1. However, the extent and type of inhibition in the presence of 0.1 M KCI for the wild type and D362E mutant is shown in Figs. 3 and 4. In addition the kinetic values are summarised in Table 1. The K i for carboxymethyl-CoA is increased 100-fold compared to the
73
W.-J. Man et al. / Biochimica et Biophysica Acta 1250 (1995) 69-75 150
charged carboxymethyl-CoA and the larger active site glutamate at the active site as a result of mutagenesis.
A 125 -
3.3. The effect of pH on the binding of carboxymethyl-CoA to wild- type E. coli citrate synthase
100 I
75
50
25
0
I
t
I
5
10
15
20
AcetyI-CoA Concentration (xlO -s pM-I) 350
B
300 250 200 150100 5O 0
0.00
I
I
I
0.01
0.02
0.03
Oxaloacetate
Concentration
0.04
(pM -11
Fig. 4. Double reciprocal plots of enzyme activity in the absence or presence of carboxymethyl-CoA for D362E mutant citrate synthase. Assays were as described in Fig. 3 except that 116 /zg of D362E mutant enzyme was used. ( - • - ) No ir,hibitor, ( - • - ) 750 /xM carboxymethylCoA.
wild-type enzyme and, surprisingly, the inhibition in the mutant with respect to oxaloacetate is now competitive rather than the classical uncompetitive nature seen with native citrate synthases [8]. This result would suggest that in the case of the mutant, the carboxymethyl-CoA is now binding in a manner that does not require the oxaloacetate induced conformational change to generate the acetyl-CoA binding site. The high K i may thus reflect relatively non-specific binding in file region of the active site in a manner that has little relation to normal acetyl-CoA binding. The inability of carboxymethyl-CoA to inhibit the D362E mutant activity in the absence of KC1 and the only very modest and possibly non-specific inhibition in the presence of KC1 is furthe[ evidence that the binding site in the E. coli enzyme is restricted. As a result, it is not possible to accommodate both the larger and negatively
A dominant additional structural feature that must be accounted for in the binding of carboxymethyl-CoA as compared with acetyl-CoA under normal assay conditions (pH 8) is the negative charge on the carboxylate (p K = 4). This negative charge would require a neutralising counterion in the active site on ternary complex formation. The obvious candidate for such a role is a positively charged His-264 and hence the binding affinity of this inhibitor should reflect the effect of change in pH on the protonation state of this histidine. Therefore the K i for carboxymethyl-CoA was measured in the presence of KC1 at pH 7, 8 and 9 and the values obtained were 0.65 _ 0.25; 6.8 _ 0.6 and 15 + 2.5 /zM respectively. These values demonstrate in particular a 10-fold decrease in K i as the pH is lowered from 8 to 7 is most consistent with a compulsory protonation over this pH range. Histidine protonation would be anticipated over this pH range while protonation of the active site aspartate cannot be excluded. Protonation of the carboxymethyl-CoA carboxylate has been proposed for the binding of this inhibitor to the enzyme from pig heart [17]. No such dramatic effect was observed for the binding of acetyl-CoA under the same conditions where the K m values were 74.6 _+ 9.5; 208 ± 41.5 and 180 __ 14.5 /xM for pH 7, 8 and 9, respectively. Thus it can be clearly seen that a change in assay pH has produced a very different effect on the binding of acetylCoA and carboxymethyl-CoA. The effect of pH on the ability of carboxymethyl-CoA to inhibit wild-type citrate synthase is shown in Fig. 5. Maximal inhibition is seen at pH 7 while the reduced
"~
60 ] ~ 40-
20-
0 0
20
40
60
[CarboxymethyI-CoA]
80
100
{~M)
Fig. 5. The effect of pH on the inhibition of wild-type citrate synthase by carboxymethyl-CoA. Assays were performed in the presence of 0.1 M KC1 as described in Fig. 1. ( - • - ) 0.1 M Potassium phosphate at pH 6.0, ( - = -) 0.1 M potassium phosphate at pH 7.0, ( - 0 - ) 0.1 M Tris-HC1 at pH 8.0, and (-*-) 0.1 M Tris-HCl at pH 9.0.
74
W.-J. Man et al. /Biochimica et Biophysica Acta 1250 (1995) 69-75
Table 2 Kinetic constants for the binding of carboxymethyl-CoAto wild-type and mutant citrate synthases Enzyme Ki (/xM) oxaloacetatevaried K i (p~M) acetyl-CoAvaried Wild-type D362E mutant
- KCl
+ KCI
- KCI
14.7 + 0.15 (UC) -
6.8 + 0.6 (UC) 620 _+14 (C)
25 _+5 (C) -
KCI 5 _+0.5 (C) 650 _+ 13 (C) +
Assay conditions were as described in Fig. 3 and Fig. 4 except that assays were also performedin the absence of KC1. UC, uncompetitiveinhibition; C, competitive inhibition.
inhibitory effect of carboxymethyl-CoA at pH 6 is parallelled by low catalytic activity at this pH. 3.4. The effect o f carboxymethyl-CoA on N A D H binding to wild-type and D362E citrate synthase
The ability of carboxymethyl-CoA to activate the D362E mutant enzyme suggested a possible interaction of the ligand with the NADH binding site. However, NADH binding studies involving fluorescence enhancement on coenzyme binding for both wild-type and mutant revealed no significant effect of the ligand on NADH binding when performed in the presence of 0.15 mM and 0.75 mM carboxymethyl-CoA respectively (data not shown). Therefore, the activation observed with carboxymethyl-CoA and the mutant enzyme in the absence of KC1, although most readily explained by a stabilisation of the R-state of the enzyme, does not involve interaction with the NADH binding site.
4. Discussion
The primary aim of the work described in this paper was to rationalise the apparent contradiction in the results obtained from the pig heart and E. coli enzyme in terms of the role of the conserved catalytic aspartate in the enzyme mechanism [2-6]. Because it is the role of the aspartate in the deprotonation of acetyl-CoA that is at issue, a method for directly assessing this step in both wild-type and the mutant (D362E) was required. The ability of the pig heart enzyme to catalyse the exchange of the methylene protons of propionyl-CoA at a high rate appeared to provide an excellent method for allowing the study of this exchange reaction in E. coli. In the event, the results in this paper highlight the fact that the E. coli enzyme is unable to catalyse this exchange, even though propionyl-CoA binding is not significantly affected. These results indicate that there are significant differences in the active site structure and function of the pig heart and E. coli enzymes. The conclusion of a more restricted active site for the E. coli enzyme was further re-enforced as a result of studies involving carboxymethyl-CoA, a presumptive tran-
sition state analogue of acetyl-CoA. Not only does this ligand have a lower affinity for the E. coli than the pig heart enzyme, but also binding to the D362E mutant is dramatically reduced, whereas acetyl-CoA binding to this mutant is not significantly affected. The authenticity of carboxymethyl-CoA as a transition-state analogue of acetyl-CoA has been questioned [16]. The lack of inhibition that is observed using the E. coli enzyme is consistent with that criticism. Since the completion of this work it has been reported that with the corresponding mutation of the pig heart enzyme (D375E) the binding affinity for carboxymethylCoA is reduced by about 100-fold [17], a reduction in line with that seen for the E. coli enzyme in the presence of KCI (Table 2). In addition, NMR studies on the binding of carboxymethyl-CoA to the E. coli enzyme in the absence of KC1 suggest that this binding does not reflect that of a reactive intermediate complex as is proposed for the pig heart enzyme [ 18]. In summary, the results suggest the active site of E. coli citrate synthase is more restricted in comparison to the pig heart enzyme based on the inability of the D362E mutant to bind carboxymethyl-CoA while no exchange of the protons of propionyl-CoA is observed with the E. coli enzyme. The ability of the pig heart enzyme to labialise both methylene protons of propionyl-CoA and to produce both stereoisomers of methylcitrate as products must be a direct reflection of binding flexibility. Moreover, the proportion of the two stereoisomers could be altered by the temperature of the enzyme assay [7] and this may be due to conformational changes allowing different substrate orientation at the active site. This result illustrates the possible danger of interpreting information derived from studies involving inhibitors or substrate analogues in terms of catalytic mechanism. The conformation adopted by the enzyme to bind a particular ligand may not reflect that required to achieve catalysis with normal substrates. In addition, a potentially rate limiting conformational change between the condensation and hydrolysis steps during catalysis has been proposed to explain the effect of limited proteolysis [19] and also mutagenesis [20] on the overall catalytic reaction of two citrate synthases. These results further support the important role of changes in protein structure [21] during catalysis by this remarkable enzyme.
w.-J. Man et al. / Biochimica et Biophysica Acta 1250 (1995) 69-75
Acknowledgements W e thank Professor 13. K o s h l a n d for kindly supplying E. coli strain D E K 15. A ' S c i e n c e and E n g i n e e r i n g Re-
search C o u n c i l ' acknowledged.
postgraduate
studentship
to W . J . M .
is
References [1] Remington, S., Wiegand, G. and Huber, R. (1982) J. Mol. Biol. 158, 111-152. [2] Karpusas, M., Branchaud, B. and Remington, S.J. (1990) Biochemistry 29, 2213-2219. [3] Alter, G., Casazza, J., Zhi, W., Nemeth, P., Srere, P.A. and Evans, C.T. (1990) Biochemistry 29, 7557-7563. [4] Karpusas, M., Holland, D. and Remington, S.J. (1991) Biochemistry 30, 6024-6031. [5] Usher, K.C., Remington, S.J., Martin, D.P. and Drueckhammer, D.G. (1994) Biochemistry 33, 7753-7759. [6] Man, W.-J., Li, Y., O'Connor, C.D. and Wilton, D.C. (1991) Biochem. J. 280, 521-526. [7] Weidman, S.W. and Drysdale, G.R. (1979) Biochem. J. 177, 169174.
75
[8] Bayer, E., Bauer, B. and Eggerer, H. (1981) Eur. J. Biochem. 120, 155-160. [9] Ner, S.S., Bhayana, V., Bell, A., Giles, I.G., Duckworth, H.W. and Bloxham, D.P. (1983) Biochemistry 22, 5243-5249. [10] Kunkel, T.A. (1985) Proc. Natl. Acad. Sci. USA 82, 488-492. [11] Taylor, I., Harrison, J.L., Timmis, K.N. and O'Connor, C.D. (1990) Mol. Microbiol. 4, 1259-1268. [12] Handford, P.A., Ner, S.S., Bloxham, D.P. and Wilton, D.C. (1988) Biochim. Biophys. Acta 953, 232-240. [13] Faloona, G.R. and Srere, P.A. (1969) Biochemistry 8, 4497-4503. [14] Duckworth, H.W. and Tong, E.K. (1976) Biochemistry 15, 108-114. [15] Man, W.-J., O'Connor, C.D. and Wilton, D.C. (1992) Biochem. Soc. Trans. 20, 291. [16] Kurz, L.C., Shah, S., Crane, B.R., Donald, L.J., Duckworth, H.W. and Drysdale, G.R. (1992) Biochemistry 31, 7899-7907. [17] Lill, U., Kollmann-Koch, A., Bibinger, A. and Eggerer, H. (1991) Eur. J. Biochem. 198, 767-773. [18] Kurz, L.C., Drysdale, G.R., Riley, M.C., Evans, C.T. and Srere, P.A. (1992) Biochemistry 31, 7908-7914. [19] Lill, U., Lefrank, S., Henschen, A. and Eggerer, H. (1992) Eur. J. Biochem. 208, 459-466. [20] Man, W.-J, Li, Y., O'Connor, C.D. and Wilton, D.C. (1994) Biochem. J. 300, 765-770. [21] Remington, S.J. (1992) Curr. Top. Cell. Regul. 33, 209-229.
Biochi~ic~a et Biophysica ~ t a Biochimicaet BiophysicaActa 1250 (1995) 69-75
The binding of propionyl-CoA and carboxymethyl-CoA to Escherichia coli citrate synthase Wai-Jin Man, Yan Li, C. David O'Connor, David C. Wilton * Departmentof Biochemistry, Universityof Southampton,BassettCrescentEast, SouthamptonS09 3TU, UK Received 4 November1994; revised 13 February 1995; accepted7 March 1995
Abstract The interaction of propionyl-CoA and acetyl-CoA with E. coli citrate synthase has been studied in order to gain insight into the structural requirements for sabstrate binding by this enzyme. In contrast to the enzyme from pig heart, the E. coli enzyme was unable to catalyse significant exchange of the methylene protons of propionyl-CoA while overall activity was very low with this enzyme. Carboxymethyl-CoA is a presumptive transition state analogue of acetyl-CoA using pig heart citrate synthase. The effect of carboxymethyl-CoA on both the native enzyme from E. coli and a catalytically active aspartate mutant (D362E) was investigated. Whereas the native enzyme was inhibited by carboxymethyl-CoA, the mutant enzyme (D362E) shows either no inhibition or minimal inhibition depending on the assay conditions. The binding of acetyl-CoA is not inhibited as a result of the mutation. The results with propionyl-CoA and carboxymethyl-CoA suggest that the active site of the E. coli enzyme is more restricted as compared with the enzyme from pig heart and, in the case of propionyl-CoA, this restriction prevents the formation of a catalytically productive enzyme-substrate complex.
Keywords: Citrate synthase; Propionyl-CoA;Carboxymethyl-CoA;Mutagenesis
1. Introduction
Citrate synthase belongs to an important group of enzymes that catalyse the formation of a new carbon-carbon bond. The enzyme catalyses the following Claisen condensation reaction: acetyl - CoA + oxaloace~:ate + H 2 0 ~ citrate + CoA + H ÷ The reaction can be divided into at least 2 steps, namely the condensation reaction to produce enzyme bound citrylCoA and the hydrolysis of citryl-CoA to produce products. Deprotonation of acetyl-CoA is a crucial event in the initial condensation reaction. Oxaloacetate binds first causing a conformational change which results in the formation of a well defined binding site for acetyl-CoA allowing subsequent attack on the si-face of oxaloacetate to form S-citryl-CoA. Most studies on citrate synthase have been carried out with the pig heart enzyme for which the crystal structure
* Corresponding author. Fax: +44 703 594459. 0167-4838/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0167-4 83 8(95)0004,~.-5
was determined by Huber's laboratory [1]. However, the identified active site residues in this enzyme are conserved in all species of citrate synthase so far studied indicating that a common mechanism is involved. A crystal structure of the enzyme from pig heart involving a ternary complex with oxaloacetate and carboxymethyl-CoA, a presumptive transition state analogue of the enol form of acetyl-CoA, allowed the proposal of an attractive mechanism for the condensation reaction [2] involving specific active-site residues. In this mechanism the conserved active-site aspartate (Asp-375) acts as a base to deprotonate the methyl carbon acetyl-CoA. This process is facilitated by the active site histidine (His-274) protonating the carbonyl oxygen to produce the enol form of acetyl-CoA [2]. This proposal was supported by mutagenesis studies [3] and further crystallographic studies involving ternary complexes with D- or L-malate and acetyl-CoA [4]. Recently, additional support for the mechanism has come from crystallographic studies using the acetyl-CoA analogues, carboxymethyldethia CoA and amidocarboxymethyldethia CoA [5].
70
W.-J. Man et al. / Biochimica et Biophysica Acta 1250 (1995) 69-75
It is apparent that much of the experimental basis for the proposed mechanism has been derived from the crystal structures of CoA analogues bound to the pig heart enzyme. Although a crystal structure of the E. coli enzyme is not available, mutagenesis studies by us involving the conserved active site aspartate suggested a different mechanism [6]. This mechanism did not involve the catalytic aspartate in the condensation stage of the reaction. Because the mechanism proposed for the pig heart enzyme was based initially on crystallographic studies using carboxymethyl-CoA, we have investigated the binding of this acetyl-CoA analogue to both wild-type and the aspartate mutant (D362E) of E. coli citrate synthase. In addition, another acetyl-CoA analogue, propionyl-CoA, can be metabolised by the pig heart enzyme and this is accompanied by rapid exchange of the ce-hydrogens of this substrate [7]. This exchange reaction should provide a method for measuring substrate deprotonation separate from overall catalysis in the E. coli enzyme. It is the mechanism of this deprotonation step that requires clarification using the E. coli enzyme. Overall, the results using the E. coli enzyme indicated that the binding site for acetyl-CoA with this citrate synthase is more catalytically restricted than that found in the pig heart enzyme. In particular, no detectable proton exchange using propionyl-CoA could be detected using the enzyme from E. coli and therefore it was not possible to investigate the role of the active site aspartate in the process. In view of these differences between the two enzymes, it must remain a possibility that mechanisms proposed for the E. coli enzyme may not be applicable to eukaryotic citrate synthases.
2. Materials and methods
2.1. Materials Carboxymethyl-CoA was prepared as described by Bayer et al. [8]. All other materials were obtained or prepared as described previously [6].
2.2. Site-directed mutagenesis The 2.1 kb SacI-SalI gltA gene encoding E. coli citrate synthase from plasmid pDB2 [9] was cloned into M13mpl9. Mutagenesis was carried according to the Kunkel method [10] using the mutagenic primer, mpl (D362E): 5'-CCGAACGTCGAATTCTACTCT-3' which was purified as previously described [11]. The bacteriophage containing the mutated gltA gene was re-sequenced with the use of specific primers before recloning the gene, as an SacI-BgllI fragment, back into pDB2. E. coli strain DEKI5, which is recombination deficient and carries a deletion of the gltA gene, was employed to harvest the mutated pDB2 plasmid.
2.3. Purification of wild-type and mutant citrate synthases Expression and purification of these enzymes was carried out as described previously [6].
2.4. Enzyme assays The 5,5'-dithio-bis(2-nitrobenzoate) (DTNB) assay of citrate synthase and NADH binding assays have been described previously [6,12]. The E. coli enzyme demonstrates allosteric properties and may be activated from Tto R-form by 0.1 M KC1 which enhances acetyl-CoA and NADH binding [13,14]. The effect of salt concentration and NADH binding of the active-site D362E mutant, has been investigated [15] and the results suggest a small shift in the equilibrium of the D362E mutant in favour of the R-state. Therefore, assays involving the binding of carboxymethyl-CoA and NADH were performed in both the absence and presence of 0.1 M KC1. All spectrophotometric assays were performed at room temperature in a total volume of 0.8 ml using an Hitachi U2000 spectrophotometer. NADH binding studies were performed using an Hitachi F2000 fluorimeter. •
3
2.5. Preparanon of H-labelled propzonyl-CoA with citrate synthase and oxaloacetate Propionyl-CoA (5 /zmol) was incubated for 40 min in 3H20 (100 mCi) in the presence of oxaloacetate (5 ~mol) and pig heart citrate synthase (0.1 mg). A total volume of 1 ml was used. The reaction mixture was freeze-dried, the residue was dissolved in dilute HCI and freeze-dried again. The residue was redissolved in 3 mM HC1 and loaded onto a DEAE-Sephacel column (1 × 25 cm), equilibrated with 3 mM HCI. Propionyl-CoA was eluted with a linear gradient formed between 150 ml of 20 mM LiC1 in 3 mM HC1 and 150 ml of 150 mM LiC1 in 3 mM HC1. 2 ml fractions were collected and monitored at 257 nm for the presence of the coenzyme A moiety and tritiated material was detected using scintillation counting. The peak corresponding to [3H]propionyl-CoA was isolated and freeze-dried. The specific activity of the propionyl-CoA was 5 . 1 . 105 dpm//zmol.
2.6. [3H]Propionyl-CoA as a substrate for wild-type and mutant E. coli citrate synthase 10-20 /zg of wild-type or D362E mutant citrate synthase was incubation in the presence of 1 mM propionylCoA, 30000 dpm of [3H]propionyl-CoA, 1 mM oxaloacetate, 100 /xM DTNB and Tris-HCl buffer (pH 8) for 1-120 min. The reactions were monitored spectrophotometrically at 412 nm and terminated using 0.05 ml of 1 M KOH and 0.01 ml of 1 M sodium acetate. The reaction mixtures were lyophilised and the radioactivity in 0.2 ml of the 3H-labelled water was determined.
71
W.-J. Man et aL / Biochimica et Biophysica Acta 1250 (1995) 69-75 100
2.7. [ 3H]Propionyl-CoA as an inhibitor for wild-type and mutant E. coli citrate synthase
The inhibitory effect of propionyl-CoA on wild-type and the D362E mutant citrate synthase was studied. The normal DTNB assay was used with the addition of 500 /zM propionyl-CoA. The J~'i for propionyl-CoA was determined by varying the concentration of oxaloacetate (0-200 #M). The reaction was followed spectrophotometrically at 412 nm using the Hitachi U2000 spectrophotometer.
~
8o
~,
60
E ~
40
2() A 0
3. Results
i
i
i
p
20
40
60
80
100
[ C a r b o x y m e t h y I - C o A ] (~M)
3.1. Propionyl-CoA as a substrate for citrate synthase from pig heart and E. coLi
150 125
In the presence of oxaloacetate pig heart citrate synthase is able to bind propionyl-CoA with high affinity and is able to catalyse the overall reaction to methyl citrate at a rate of about 0.01% of that observed for citrate formation [7]. In addition, the enzyme is able to rapidly exchange the c~-protons of propionyl-CoA with the pro-S hydrogen being exchanged 15 times faster than the pro-R hydrogen. This facile exchange catalysed by the enzyme in the presence of oxaloacetate provided a potential system for investigating the role of the active site aspartate in substrate deprotonation in the case of the E. coli enzyme. Therefore, [2-3H]propionyl-CoA was prepared by making use of the exchange reaction catalysed by pig heart enzyme and the ability of the E. coli enzyme to catalyse this exchange and the overall reaction was investigated. The results of studies using propionyI-CoA and both the wild-type and D362E mutant citrate synthase are shown in Table 1. Under the assay conditions there was a very low but detectable overall activity suggesting formation of methyl citrate in the wild-type enzyme. However, no such activity could be detected in the D362E mutant consistent with the observation that this mutant has only about 1% of the overall activity of the wild-type enzyme [6]. The propionyl-CoA activity s~een with the wild-type enzyme was proportionally similar to that observed with the pig heart enzyme. However, when the ability of the wild-type
~
A
100
~
75
E
50
N C 1.1.1
25 0 20
40
60
80
100
[ C a r b o x y r n e t h y I - C o A ] (pM}
Fig. 1. The effect of carboxymethyl-CoA on the activity of wild-type and mutant citrate synthases. The reactions were performed in the presence of 125/zM acetyl-CoA, 125 /.tM oxaloacetate and 125/zM DTNB in 0.1 M Tris-HC1 buffer (pH 8.0). Assays were performed in the presence (A) and absence (B) of 0.1 M KC1 and activities are expressed as a percentage of control activity in the absence of carboxymethyl-CoA. (-O-) Wild-type enzyme (4.4 p.g), and ( - • - ) D362E mutant enzyme (116 /zg).
E. coli enzyme to release tritium from [2-3H]propionylCoA was studied no significant exchange of tritium could be detected even with large amounts of enzyme. It should be emphasised that the compulsory loss of tritium associated with the very low rate of propionyl-CoA metabolism would not be detected in these experiments using the E. coli enzyme. Under identical conditions there was a considerable release of tritium by the pig heart enzyme. The
Table 1 Propionyl-CoA as the substrate or inhibitor for E. coli citrate synthase Citrate synthase
Activity (U/mg)
Tritium exchange ( / x m o l / m i n per mg)
Propionyl CoA K i (/zM)
Pig heart Wild-type E. coli D362E E. coli
0.002 + 0.0005 < 0.0003 -
3.6 + 0.3 < 0.025 < 0.004
540 + 200 670 + 170
All assays were performed in 0.1 M Tris-HCl buffer at pH 8.0. Assays for enzyme activity with propionyl-CoA as substrate were performed using 250 poM propionyl-CoA and 500 p.M oxaloacetate. Inhibition studies were performed using 250 /.LM acetyl-CoA, 500 /.LM and the concentration of oxaloacetate was varied from 0 - 2 0 0 / x M . 4 ~ g of wild-type and 116/zg D362E mutant E. coli citrate synthase was used. Tritium exchange studies were performed as described in Section 2.
72
W.-J. Man et al. / Biochimica et Biophysica Acta 1250 (1995) 69-75
rate of tritium release for the pig heart enzyme (3.6 /zmol/min per mg) compares favourably with the value of 6.6 ~ m o l / m i n per mg previously reported for this enzyme [7]. In view of the inability of the E. coli enzyme to use propionyl-CoA as a substrate it was of interest to assess the binding of the propionyl-CoA to the enzyme. Since the actual rate observed with the wild-type enzyme was only just within the limit of sensitivity of the normal DTNB-linked assay for this enzyme, binding was assessed in terms of the ability of the propionyl-CoA to act as a competitive inhibitor. The K i values for wild-type and D362E mutant E. coli citrate synthase were 540 /~M and 667 /zM, respectively, indicating that it does bind weakly to the E. coli enzyme. The corresponding K m value for acetyl-CoA under these conditions was 350 /xM [6]. Therefore, the lack of activity exhibited with the E. coli enzyme cannot be explained by the inability of propionyl-CoA to bind to the enzyme.
3.2. The effect of carboxymethyl-CoA on the activity of E. coli citrate synthase The effect of varying concentrations of carboxymethylCoA on the overall activity of citrate synthase wild-type and D362E mutant enzymes was measured under standard conditions in the absence or presence of 0.1 M KCI. The presence of 0.1 M KC1 favours the R-form of the enzyme with an increased affinity for acetyl-CoA [13,14] (see Section 2). In the wild-type enzyme the inhibitory effect of carboxymethyl-CoA is more apparent in the presence of KC1 (Fig. 1A) as compared with the absence of KC1 (Fig. 1B) suggesting that carboxymethyl-CoA binds more avidly to the R-form of the enzyme as is known for acetyl-CoA binding [13]. However, a dramatic effect is seen when the action of carboxymethyl-CoA on the mutant enzyme is
300
A 250
200
150 ,11
100
50
0 0
r
i
i
5
10
15
20
AcetyI-CoA Concentration (xl0 -~ pM-1l 200
B
1so
~
~
100
s0
0 0.00
~
i
i
i
0.01
0.02
0.03
0.04
Oxaloacetate Concentration (pM-1) Fig. 3. Double reciprocal plots of enzyme activity in the absence or presence of carboxymethyl-CoA for wild-type citrate synthase. Assays were performed (A) in the presence of 250 /xM oxaloacetate or (B) 250 /-~M acetyl-CoA using wild-type citrate synthase (4 ixg). All assays were performed in the presence of O. 1 M KCI. (- • - ) No inhibitor, and (- • -) 25 /xM carboxymethyl-CoA.
200
~ w
15o
I00
E N =
50-
J
i
r
,
i
100
200
300
400
500
600
[ C a r b o x y m e t h y I - C o A ] (~M) Fig. 2. The effect of high concentrations of carboxymethyl-CoA on the activity of the D362E mutant citrate synthase. The assay conditions were as described in Fig. 1 except that the concentration of carboxymethyl-CoA used was up to 600 ~ M . (- • -) Assay in the absence of KC1, and (- • -) assay in the presence of 0.1 M KCI.
examined. In the absence of KC1, carboxymethyl-CoA partially activates the enzyme, a result most readily explained by the inhibitor binding at a second site and stabilising the R-form. This conclusion is supported by assaying the effect of carboxymethyl-CoA in the presence of KC1 which will also stabilise the R-form. Under these conditions no activation is observed while significant inhibition can be detected in the presence of high concentrations of carboxymethyl-CoA (Fig. 2). It was not possible to analyze the effects of carboxymethyl-CoA on the activity of the mutant enzyme in the absence of KC1. However, the extent and type of inhibition in the presence of 0.1 M KCI for the wild type and D362E mutant is shown in Figs. 3 and 4. In addition the kinetic values are summarised in Table 1. The K i for carboxymethyl-CoA is increased 100-fold compared to the
73
W.-J. Man et al. / Biochimica et Biophysica Acta 1250 (1995) 69-75 150
charged carboxymethyl-CoA and the larger active site glutamate at the active site as a result of mutagenesis.
A 125 -
3.3. The effect of pH on the binding of carboxymethyl-CoA to wild- type E. coli citrate synthase
100 I
75
50
25
0
I
t
I
5
10
15
20
AcetyI-CoA Concentration (xlO -s pM-I) 350
B
300 250 200 150100 5O 0
0.00
I
I
I
0.01
0.02
0.03
Oxaloacetate
Concentration
0.04
(pM -11
Fig. 4. Double reciprocal plots of enzyme activity in the absence or presence of carboxymethyl-CoA for D362E mutant citrate synthase. Assays were as described in Fig. 3 except that 116 /zg of D362E mutant enzyme was used. ( - • - ) No ir,hibitor, ( - • - ) 750 /xM carboxymethylCoA.
wild-type enzyme and, surprisingly, the inhibition in the mutant with respect to oxaloacetate is now competitive rather than the classical uncompetitive nature seen with native citrate synthases [8]. This result would suggest that in the case of the mutant, the carboxymethyl-CoA is now binding in a manner that does not require the oxaloacetate induced conformational change to generate the acetyl-CoA binding site. The high K i may thus reflect relatively non-specific binding in file region of the active site in a manner that has little relation to normal acetyl-CoA binding. The inability of carboxymethyl-CoA to inhibit the D362E mutant activity in the absence of KC1 and the only very modest and possibly non-specific inhibition in the presence of KC1 is furthe[ evidence that the binding site in the E. coli enzyme is restricted. As a result, it is not possible to accommodate both the larger and negatively
A dominant additional structural feature that must be accounted for in the binding of carboxymethyl-CoA as compared with acetyl-CoA under normal assay conditions (pH 8) is the negative charge on the carboxylate (p K = 4). This negative charge would require a neutralising counterion in the active site on ternary complex formation. The obvious candidate for such a role is a positively charged His-264 and hence the binding affinity of this inhibitor should reflect the effect of change in pH on the protonation state of this histidine. Therefore the K i for carboxymethyl-CoA was measured in the presence of KC1 at pH 7, 8 and 9 and the values obtained were 0.65 _ 0.25; 6.8 _ 0.6 and 15 + 2.5 /zM respectively. These values demonstrate in particular a 10-fold decrease in K i as the pH is lowered from 8 to 7 is most consistent with a compulsory protonation over this pH range. Histidine protonation would be anticipated over this pH range while protonation of the active site aspartate cannot be excluded. Protonation of the carboxymethyl-CoA carboxylate has been proposed for the binding of this inhibitor to the enzyme from pig heart [17]. No such dramatic effect was observed for the binding of acetyl-CoA under the same conditions where the K m values were 74.6 _+ 9.5; 208 ± 41.5 and 180 __ 14.5 /xM for pH 7, 8 and 9, respectively. Thus it can be clearly seen that a change in assay pH has produced a very different effect on the binding of acetylCoA and carboxymethyl-CoA. The effect of pH on the ability of carboxymethyl-CoA to inhibit wild-type citrate synthase is shown in Fig. 5. Maximal inhibition is seen at pH 7 while the reduced
"~
60 ] ~ 40-
20-
0 0
20
40
60
[CarboxymethyI-CoA]
80
100
{~M)
Fig. 5. The effect of pH on the inhibition of wild-type citrate synthase by carboxymethyl-CoA. Assays were performed in the presence of 0.1 M KC1 as described in Fig. 1. ( - • - ) 0.1 M Potassium phosphate at pH 6.0, ( - = -) 0.1 M potassium phosphate at pH 7.0, ( - 0 - ) 0.1 M Tris-HC1 at pH 8.0, and (-*-) 0.1 M Tris-HCl at pH 9.0.
74
W.-J. Man et al. /Biochimica et Biophysica Acta 1250 (1995) 69-75
Table 2 Kinetic constants for the binding of carboxymethyl-CoAto wild-type and mutant citrate synthases Enzyme Ki (/xM) oxaloacetatevaried K i (p~M) acetyl-CoAvaried Wild-type D362E mutant
- KCl
+ KCI
- KCI
14.7 + 0.15 (UC) -
6.8 + 0.6 (UC) 620 _+14 (C)
25 _+5 (C) -
KCI 5 _+0.5 (C) 650 _+ 13 (C) +
Assay conditions were as described in Fig. 3 and Fig. 4 except that assays were also performedin the absence of KC1. UC, uncompetitiveinhibition; C, competitive inhibition.
inhibitory effect of carboxymethyl-CoA at pH 6 is parallelled by low catalytic activity at this pH. 3.4. The effect o f carboxymethyl-CoA on N A D H binding to wild-type and D362E citrate synthase
The ability of carboxymethyl-CoA to activate the D362E mutant enzyme suggested a possible interaction of the ligand with the NADH binding site. However, NADH binding studies involving fluorescence enhancement on coenzyme binding for both wild-type and mutant revealed no significant effect of the ligand on NADH binding when performed in the presence of 0.15 mM and 0.75 mM carboxymethyl-CoA respectively (data not shown). Therefore, the activation observed with carboxymethyl-CoA and the mutant enzyme in the absence of KC1, although most readily explained by a stabilisation of the R-state of the enzyme, does not involve interaction with the NADH binding site.
4. Discussion
The primary aim of the work described in this paper was to rationalise the apparent contradiction in the results obtained from the pig heart and E. coli enzyme in terms of the role of the conserved catalytic aspartate in the enzyme mechanism [2-6]. Because it is the role of the aspartate in the deprotonation of acetyl-CoA that is at issue, a method for directly assessing this step in both wild-type and the mutant (D362E) was required. The ability of the pig heart enzyme to catalyse the exchange of the methylene protons of propionyl-CoA at a high rate appeared to provide an excellent method for allowing the study of this exchange reaction in E. coli. In the event, the results in this paper highlight the fact that the E. coli enzyme is unable to catalyse this exchange, even though propionyl-CoA binding is not significantly affected. These results indicate that there are significant differences in the active site structure and function of the pig heart and E. coli enzymes. The conclusion of a more restricted active site for the E. coli enzyme was further re-enforced as a result of studies involving carboxymethyl-CoA, a presumptive tran-
sition state analogue of acetyl-CoA. Not only does this ligand have a lower affinity for the E. coli than the pig heart enzyme, but also binding to the D362E mutant is dramatically reduced, whereas acetyl-CoA binding to this mutant is not significantly affected. The authenticity of carboxymethyl-CoA as a transition-state analogue of acetyl-CoA has been questioned [16]. The lack of inhibition that is observed using the E. coli enzyme is consistent with that criticism. Since the completion of this work it has been reported that with the corresponding mutation of the pig heart enzyme (D375E) the binding affinity for carboxymethylCoA is reduced by about 100-fold [17], a reduction in line with that seen for the E. coli enzyme in the presence of KCI (Table 2). In addition, NMR studies on the binding of carboxymethyl-CoA to the E. coli enzyme in the absence of KC1 suggest that this binding does not reflect that of a reactive intermediate complex as is proposed for the pig heart enzyme [ 18]. In summary, the results suggest the active site of E. coli citrate synthase is more restricted in comparison to the pig heart enzyme based on the inability of the D362E mutant to bind carboxymethyl-CoA while no exchange of the protons of propionyl-CoA is observed with the E. coli enzyme. The ability of the pig heart enzyme to labialise both methylene protons of propionyl-CoA and to produce both stereoisomers of methylcitrate as products must be a direct reflection of binding flexibility. Moreover, the proportion of the two stereoisomers could be altered by the temperature of the enzyme assay [7] and this may be due to conformational changes allowing different substrate orientation at the active site. This result illustrates the possible danger of interpreting information derived from studies involving inhibitors or substrate analogues in terms of catalytic mechanism. The conformation adopted by the enzyme to bind a particular ligand may not reflect that required to achieve catalysis with normal substrates. In addition, a potentially rate limiting conformational change between the condensation and hydrolysis steps during catalysis has been proposed to explain the effect of limited proteolysis [19] and also mutagenesis [20] on the overall catalytic reaction of two citrate synthases. These results further support the important role of changes in protein structure [21] during catalysis by this remarkable enzyme.
w.-J. Man et al. / Biochimica et Biophysica Acta 1250 (1995) 69-75
Acknowledgements W e thank Professor 13. K o s h l a n d for kindly supplying E. coli strain D E K 15. A ' S c i e n c e and E n g i n e e r i n g Re-
search C o u n c i l ' acknowledged.
postgraduate
studentship
to W . J . M .
is
References [1] Remington, S., Wiegand, G. and Huber, R. (1982) J. Mol. Biol. 158, 111-152. [2] Karpusas, M., Branchaud, B. and Remington, S.J. (1990) Biochemistry 29, 2213-2219. [3] Alter, G., Casazza, J., Zhi, W., Nemeth, P., Srere, P.A. and Evans, C.T. (1990) Biochemistry 29, 7557-7563. [4] Karpusas, M., Holland, D. and Remington, S.J. (1991) Biochemistry 30, 6024-6031. [5] Usher, K.C., Remington, S.J., Martin, D.P. and Drueckhammer, D.G. (1994) Biochemistry 33, 7753-7759. [6] Man, W.-J., Li, Y., O'Connor, C.D. and Wilton, D.C. (1991) Biochem. J. 280, 521-526. [7] Weidman, S.W. and Drysdale, G.R. (1979) Biochem. J. 177, 169174.
75
[8] Bayer, E., Bauer, B. and Eggerer, H. (1981) Eur. J. Biochem. 120, 155-160. [9] Ner, S.S., Bhayana, V., Bell, A., Giles, I.G., Duckworth, H.W. and Bloxham, D.P. (1983) Biochemistry 22, 5243-5249. [10] Kunkel, T.A. (1985) Proc. Natl. Acad. Sci. USA 82, 488-492. [11] Taylor, I., Harrison, J.L., Timmis, K.N. and O'Connor, C.D. (1990) Mol. Microbiol. 4, 1259-1268. [12] Handford, P.A., Ner, S.S., Bloxham, D.P. and Wilton, D.C. (1988) Biochim. Biophys. Acta 953, 232-240. [13] Faloona, G.R. and Srere, P.A. (1969) Biochemistry 8, 4497-4503. [14] Duckworth, H.W. and Tong, E.K. (1976) Biochemistry 15, 108-114. [15] Man, W.-J., O'Connor, C.D. and Wilton, D.C. (1992) Biochem. Soc. Trans. 20, 291. [16] Kurz, L.C., Shah, S., Crane, B.R., Donald, L.J., Duckworth, H.W. and Drysdale, G.R. (1992) Biochemistry 31, 7899-7907. [17] Lill, U., Kollmann-Koch, A., Bibinger, A. and Eggerer, H. (1991) Eur. J. Biochem. 198, 767-773. [18] Kurz, L.C., Drysdale, G.R., Riley, M.C., Evans, C.T. and Srere, P.A. (1992) Biochemistry 31, 7908-7914. [19] Lill, U., Lefrank, S., Henschen, A. and Eggerer, H. (1992) Eur. J. Biochem. 208, 459-466. [20] Man, W.-J, Li, Y., O'Connor, C.D. and Wilton, D.C. (1994) Biochem. J. 300, 765-770. [21] Remington, S.J. (1992) Curr. Top. Cell. Regul. 33, 209-229.