On the thermodynamic equilibrium between (R)-2-hydroxyacyl-CoA and 2-enoyl-CoA Anutthaman Parthasarathy 1 , Wolfgang Buckel 1 and David M. Smith 2 1 Laboratory for Microbiology, Philipps-Universita ¨ t, Marburg, Germany 2 Centre for Computational Solutions in the Life Sciences, Rudjer Boskovic Institute, Zagreb, Croatia Introduction Dehydratases catalyze a,b-eliminations of water from hydroxy compounds, and form a large class of enzymes; over 100 different types are listed in the enzyme nomenclature database (EC 4.2.1.–). In most cases, the hydroxyl group is located in the b-position of an adjacent carboxylate, CoA-thioester or ketone, and the a-proton to be removed is thus activated. In a-amino acid fermentation pathways, however, dehy- dratases are found, whose substrates are a-hydroxya- cyl-CoA derivatives [1,2]. In such cases, the b-hydrogen to be removed during a,b-elimination of water has an approximate pK a of 40. The requisite activation of this proton is achieved by transient addition of one high-energy electron to the thioester carbonyl, forming a ketyl radical anion (3, Scheme 1). This allows the elimination of the a-hydroxyl group [3,4] (3 fi 4) and lowers the pK a of the b-hydrogen in the resulting enoxy radical intermediate (4)byat least 26 units [5]. Recycling of the initiatory electron from the second ketyl intermediate thus produced (5) yields enoyl-CoA (2) and completes the catalytic cycle. Recently, the mechanism shown in Scheme 1 has received strong support with the reported observation of the allylic ketyl radical intermediate (5) in the enyzmatic dehydration of (R)-2-hydroxyisocaproyl- CoA [1a, R = CH(CH 3 ) 2 ; Fig. 1].[4] However, no such observations have yet been possible for the analogous dehydrations of (R)-2-hydroxyglutaryl-CoA (1b,R=CH 2 CO 2 H) or (R)-lactyl-CoA (1c, R = H). Whereas the equilibrium constants (K )ofb-hydrox- yacyl-CoA, c-hydroxyacyl-CoA or d-hydroxyacyl-CoA typically lie around unity [6], the situation is much less clear for the a-hydroxyacyl-CoA derivatives. For Keywords ab initio calculations; enzymes; kinetics; solvent effects; substituent effects Correspondence D. M. Smith, Centre for Computational Solutions in the Life Sciences, Rudjer Boskovic Institute, Bijenicka 54, 10000 Zagreb, Croatia Fax: +385 1 456 1182 Tel: +385 1 456 1182 E-mail: David.Smith@irb.hr (Received 20 December 2009, revised 25 January 2010, accepted 28 January 2010) doi:10.1111/j.1742-4658.2010.07597.x A combined experimental and computational approach has been applied to investigate the equilibria between several a-hydroxyacyl-CoA compounds and their 2-enoyl-CoA derivatives. In contrast to those of their b, c and d counterparts, the equilibria for the a-compounds are relatively poorly char- acterized, but qualitatively they appear to be unusually sensitive to substit- uents. Using a variety of techniques, we have succeeded in measuring the equilibrium constants for the reactions beginning from 2-hydroxyglutaryl- CoA and lactyl-CoA. A complementary computational evaluation of the equilibrium constants shows quantitative agreement with the measured values. By examining the computational results, we arrive at an explanation of the substituent sensitivity and provide a prediction for the, as yet unmeasured, equilibrium involving 2-hydroxyisocaproyl-CoA. Abbreviations Nbs 2 , 5,5¢-dithiobis(2-nitrobenzoate); TFA, trifluoroacetic acid; THF, trifluoroacetic acid. 1738 FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS example, it has recently been determined that the enzy- matic dehydration of 1a (also known as 2-hydroxy- 4-methylpentanoyl-CoA), derived from the amino acid (S)-leucine, to isocaprenoyl-CoA (4-methyl-2-pente- noyl-CoA, probably the E-isomer, 2a), occurs irrevers- ibly, within the limits of detection [7]. In contrast, the equilibrium of the dehydration of (R)-lactyl-CoA to acryloyl-CoA (1c fi 2c) strongly favors the hydroxy compound. Under physiological conditions, (R)-lactyl- CoA is only effectively dehydrated, because the very small equilibrium concentration of the unsaturated compound (2c) is irreversibly trapped by the consecu- tive reductase, resulting in propionyl-CoA [8,9]. With (R)-2-hydroxyglutaryl-CoA (1b) and (E)-glutaconyl- CoA (2b) as substrates, the equilibrium appears to lie more in the middle [10], although the value of K b , like those of K a and K c , is presently unknown. In order to place the known qualitative results on a more solid footing, we adopted a combined experimen- tal and computational approach to characterize the equilibria shown in Fig. 1. Specifically, using a bidirec- tional kinetic analysis, we present experimental deter- minations of K b and K c . The obtained values are supported and rationalized by high-level ab initio molecular orbital calculations, which are also used to obtain a value for the elusive K a . Results and Discussion For the determination of the experimental value of K b , we performed activity measurements in both the for- ward and reverse directions, requiring the preparation of both 1b and 2b . Although an enzymatic method was potentially applicable for the production of 1b, the incubations in that case also yielded the thermo- dynamically more favorable (R)-4-hydroxyglutaryl- CoA, which is not dehydrated [11]. To avoid this complication, a chemical method for the preparation of 1b was used; namely, (R)-2-hydroxyglutaryl-CoA (1b) was prepared by the direct reaction of CoASH and (R)-butyrolactone-5-carboxylchloride in aqueous NaHCO 3 [11]. After the thiol had been consumed, the reaction was acidified to pH 0.5 and incubated at 25 °C for 3 h until the equilibrium between lactone- CoA and 2-hydroxyglutaryl-CoA (1 : 4) was estab- lished. Prior to use, the mixture was neutralized, whereby equilibration was terminated. It was shown that the remaining lactone-CoA did not interfere with the dehydration. In the case of glutaconyl-CoA (2b) preparation, there were no complicating factors, and the target compound was prepared by incubation of glutaconate with acetyl-CoA, catalyzed by glutaconate- CoA transferase from Acidaminococcus fermentans produced in Escherichia coli [12]. For our source of 2-hydroxyglutaryl-CoA dehydra- tase, we employed the strictly anaerobic bacterium Clostridium symbiosum, which is involved in the fermentation of glutamate to ammonia, CO 2 , acetate, butyrate, and H 2 . The heterodimeric enzyme O SCoA R O SCoA R H H 12 OH H H H H 2 O 2-hydroxyisocaproyl-CoA (1a) 2-hydroxyglutaryl-CoA (1b) lactyl-CoA (1c) E-isocaprenoyl-CoA (2a) + H 2 O E-glutaconyl-CoA (2b) + H 2 O acryloyl-CoA (2c) + H 2 O K conc. = [2] [H 2 O] [1] [2] [1] K = K R=CH(CH 3 ) 2 R=CH 2 CO 2 H R=H K a >> 1 K b ~ 1 K c << 1 K conc. 55.5 = Fig. 1. Dehydrations of the a-hydroxyacyl- CoA derivatives discussed in this work. As the biochemical experiments were performed in dilute aqueous conditions, we employ K, as opposed to K conc. , for convenience. Scheme 1. Electron recycling mechanism for the conversion of a-hydroxyacyl-CoA derivatives to the corresponding enoyl-CoA compounds. A. Parthasarathy et al. Equilibrium between hydroxyacyl-CoA and enoyl-CoA FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS 1739 (48 + 43 kDa) contains, per mole (91 kDa), 1 mol of FMNH 2 and 8 mol of iron + 8 mol of sulfur, proba- bly as two [4Fe–4S] clusters. It has to be activated by incubation with a reducing agent, ATP, and Mg 2+ , mediated by a homodimeric protein (2 · 27 kDa) with one [4Fe–4S] + cluster between the two subunits. Thereby, one electron is transferred from the activator protein to the dehydratase, driven by hydrolysis of two molecules of ATP. Prior to use in our activity mea- surements, 2-hydroxyglutaryl-CoA dehydratase, puri- fied from C. symbiosum, was activated under strict anaerobic conditions with ATP, MgCl 2 and dithionite, mediated by catalytic amounts of the activator from A. fermentans produced in E. coli [13]. The dehydration (1b fi 2b) and hydration (2b fi 1b) reactions were followed spectrophotometri- cally at 290 nm (e 290 nm = 2.1 mm )1 Æcm )1 ). Although the absorbance maximum of enoyl-CoA (2b) lies at 260 nm (De 260 nm = 6.0 mm )1 Æcm )1 ), the longer wave- length was chosen to avoid interference with the high absorbance of the adenine moiety of CoA (e 260 nm =16mm )1 Æcm )1 ) [7]. The activity measurements between 0 and 1 mm (R)- 2-hydroxyglutaryl-CoA (1b) and 0 and 5 mm glutaco- nyl-CoA (2b) gave smooth Michaelis–Menten curves, from which K m and k cat values could be calculated by simulation. In experiments starting with (R)-2-hydroxyglutaryl-CoA (1b), we obtained a K m of 0.052 ± 0.003 mm and a k cat of 83 ± 8 s )1 , resulting in a specificity constant [k cat ⁄ K m (1b)] of 1600 ± 300 s )1 Æmm )1 . Beginning the reaction with, instead, (E)-glutaconyl-CoA (2b) resulted in a K m of 0.25 ± 0.02 mm and a k cat of 7.0 ± 0.7 s )1 , associated with a specificity constant [k cat ⁄ K m (2b)] of 28±6s )1 Æmm )1 . K b was subsequently calculated from the Briggs–Haldane equation: K b ¼½k cat =K m ð1bÞ=½k cat =K m ð2bÞ ¼ 1600=28 ¼ 57 Æ 1:5 For the measurement of K c , the equilibrium constant for (R)-lactyl-CoA (1c) and acryloyl-CoA (2c), we purified lactyl-CoA dehydratase from the strict anaer- obe Clostridium propionicum to apparent homogeneity. Like 2-hydroxyglutaryl-CoA dehydratase, lactyl-CoA dehydratase is a heterodimer (a, 48 kDa; b, 41 kDa) containing two [4Fe–4S] clusters and substoichiometric amounts of FMN and riboflavin [9,14]. The purified enzyme was treated with 3-pentynoyl-CoA to abolish a slight acryloyl-CoA reductase activity [9,15]. Activation of the dehydratase occurred under conditions similar to that used for the C. symbiosum dehydratase in the presence of Mg-ATP, dithionite, and the activator from A. fermentans [13]. The CoA-thioesters were pre- pared from acrylate, (R)-lactate and 3-pentynoic acid [16] by the carbonyl-diimidazole method [17], and analyzed enzymatically and by MALDI-TOF MS [18]. The kinetic parameters for the hydration of acryloyl-CoA (2c)to(R)-lactyl-CoA (1c) were measured as K m = 0.150 ± 0.004 mm and V max = 85 ± 6 UÆmg )1 , yielding k cat =126±10s )1 and [k cat ⁄ K m (2c)] = 0.84 ± 0.05 · 10 6 s )1 Æm )1 . Because the equilibrium concentration of acryloyl-CoA (2c)is very low, it was more difficult to determine the kinetics of (R)-lactyl-CoA dehydration (1c fi 2c). Reasonable estimates are given by: K m = 0.32 ± 0.02 mm and V max = 3.0 ± 0.4 UÆmg )1 , with k cat = 4.5 ± 0.6 s )1 and [k cat ⁄ K m (1c)] = 1.41 ± 0.1 · 10 4 s )1 Æm )1 ). Substi- tution of these data into the Briggs–Haldane equation yields: K c ¼½k cat =K m ð1cÞ=½k cat =K m ð2cÞ ¼ 0:017 Æ 0:007 This low value of K c corroborates the high redox potential of the acryloyl-CoA ⁄ propionyl-CoA pair (E 0 ¢ = + 69 mV) as compared with those of the higher homologs of 2-enoyl-CoA ⁄ acyl-CoA (E 0 ¢ = )10 mV) [19]. The relative magnitudes of K b and K c confirm an unexpectedly large ($ 20 kJÆmol )1 ) substituent effect on the dehydration equilibrium, arising from the pres- ence of a carboxymethyl group on the b-carbon in 1b in place of a hydrogen atom in 1c. This fact, combined with the similarly large effect that was apparent upon further substitution by the isopropyl group in 1a, led us to perform ab initio molecular orbital calculations in order to seek an explanation. To enable high-accuracy calculations, we elected to replace the adenylphosphopantetheine chain of CoA by the S-CH 3 group, resulting in the model systems shown in Fig. 2. We expected this substitution to have only a minor effect on the individual equilibrium con- stants and, because it is adopted uniformly, there should be virtually no effect on the relative equilibrium constants. Although our final goal was to compute the free energies of the reactions shown in Fig. 1 under aqueous conditions, we elected to present the gas- phase results as well. The rationale behind this is that the gas-phase calculations encompass the fundamental electronic effects governing the differences in the equi- librium constants. By decomposing the final aqueous energy differences into a gas-phase component and a component related to solvation, we are thus able to comment on the extent to which each of these aspects influences the final result. Equilibrium between hydroxyacyl-CoA and enoyl-CoA A. Parthasarathy et al. 1740 FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS Table 1 shows the standard (referenced to 1 atm of pressure) free energy change for each reaction in the gas phase [1 (g) fi 2 (g) +H 2 O (g) ; DG  ðgÞ ]. (Although 1 and 2 strictly represent the CoA thioester, for simplic- ity, we keep the same notation for the methyl thioest- ers used in the calculations.) In the gas phase, the dehydration of 1c was found to be mildly (2.9 kJÆmol )1 ) endergonic. The introduc- tion of a carboxymethyl substituent at the b-carbon was found to preferentially stabilize the 2-enoyl spe- cies (2b), such that dehydration of 2a was exergonic by 1.6 kJÆmol )1 . The small associated substituent effect (4.5 kJÆmol )1 ) can be rationalized by the mild (net) capability of the alkyl substituent to donate electrons to the electron-deficient b-carbon in 2c. Indeed, acrylamide, with no substituent at the b-posi- tion, has been shown to act as a toxic electrophilic agent [20], an effect that should be more pronounced in acryloyl-CoA (2c). The more electron-donating iso- propyl substituent results in a larger preferential sta- bilization for the enoyl species (2a), such that the dehydration of 1a is exergonic by 9.0 kJÆmol )1 . Both of these relatively small, inherent (gas-phase) substitu- ent effects are more in line with qualitative expecta- tions than the values apparent from the measured solution-phase equilibrium constants. It thus appears that the explanation of the unusually large substituent effects is not related to fundamental electronic factors at the molecular level. The standard free energy changes in aqueous solu- tion [1 (aq) fi 2 (aq) +H 2 O (aq) ; DG Ã ðaqÞ , referenced to 1 molÆL )1 ] could be expected to preferentially favor the products, simply because of the sizeable solvation free energy of water [experimentally, DG Ã s (H 2 O) = )26.5 kJÆmol )1 [21]]. This preference for the bimo- lecular products is reduced by the reference state correction (from 1 atm to 1 molÆL )1 )ofRT ln( ~ RT) of 7.9 kJÆmol )1 (at 298 K) for each species to )18.6 kJÆmol )1 ( ~ R = 0.082053 K )1 ). In the case of (R)-lactyl-CoA, the inherent product preference in solution is partially compensated for by the relatively large (absolute) value of DG Ã s (1c) as compared with DG Ã s (2c) (Table 2). The result of these competing effects is that DG Ã ðaqÞ (C) is only 2.8 kJÆmol )1 less than the corresponding gas-phase value. Despite the poten- tial uncertainties involved, the final calculated value for DG Ã ðaqÞ (C) (0.1 kJÆmol )1 ) is in very good agreement with that derived from the measured equilibrium con- stant of 0.017 (0.1 kJÆmol )1 ). The absolute magnitude of the free energies of solvation of 1b and 2b are much larger than those of 1c and 2c, because of the presence of the hydrophilic + + + ++ 1a 1b 1c 2a 2b 2c K calc. a = 1610 K expt. a > 1000 K calc. b = 8.42 K expt. b = 57 K calc. c = 0.02 K expt. c = 0.017 Fig. 2. Comparison of the calculated and experimental equilibrium constants determined in this work. Table 1. Experimentally determined and calculated values for the equilibria represented by Fig. 1. Equilibrium Calculated [G3(MP2)] Experimental DG  ðgÞ a,b DG Ã ðaqÞ a,c K d DG Ã ðaqÞ a,c K d A )9.0 )28.3 1610 < )27.1 > 1000 B )1.6 )15.2 8.42 )19.9 57 C 2.9 0.1 0.02 0.1 0.017 a kJÆmol )1 at 298 K. b 1 atm reference. c 1 molÆL )1 reference. d Dimensionless, K = K conc. ⁄ 55.5 (see Fig. 1). Table 2. Calculated free energies of solvation (DG Ã s ) for the species shown in Fig. 1. 1a 2a 1b 2b 1c 2c H 2 O DG Ã s a )4.4 )6.8 )43.7 )40.6 )16.4 )2.5 )24.7 a The calculated free energy of solvation (in kJÆmol )1 ) for x (g) (1 molÆL -1 ) fi x (aq) (1 molÆL )1 ). The final DG Ã ðaqÞ values in Table 1 are corrected by RT ln( ~ RT) for each species. See text. A. Parthasarathy et al. Equilibrium between hydroxyacyl-CoA and enoyl-CoA FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS 1741 carboxylic acid groups. Even though the hydroxyacyl species (1b) species is again solvated more strongly than the enoyl one (2b), the difference between them, and hence the associated effect on DG Ã ðaqÞ (B), is much smaller than for reaction C. The favorable contribu- tion from DG Ã s (H 2 O) is therefore counteracted to a much smaller extent, resulting in the value for DG Ã ðaqÞ (B) being 13.6 kJÆmol )1 more negative than DG  ðgÞ (B). The corresponding calculated value of K b = 8.42 deviates somewhat from the measured value of K b = 57. In terms of energy, however, the discrepancy of only 4.7 kJÆmol )1 is certainly within acceptable limits for a solution-phase property. The more hydrophobic nature of 1a and 2a is reflected in their less favorable solvation free energies (Table 2). In this case, however, it is the enoyl species (2a) that is better solvated than the hydroxyacyl one (1a). This serves to slightly reinforce the favorable effect of DG Ã s (H 2 O) rather than to counteract it, as occurred for equilibria B and C. The final result is that equilibrium A is predicted to lie very far to the right, with associated values of DG Ã ðaqÞ (A)=)28.3 kJÆmol )1 and K a = 1610. Given the good agreement between theory and experiment obtained for equilibria B and C, we are confident that these values provide a reason- ably accurate description of the thermodynamics of 2-hydroxyisocaproyl-CoA dehydration. They are cer- tainly consistent with the fact that the equilibrium concentration of 1a was not detectable in experiments concerning its conversion into 2a [7]. The precise experimental determination of K a , however, still remains a challenge for the future. Conclusion In summary, and in agreement with previous qualita- tive observations, the measurements presented here confirm an unusually large effect of the substituent at the b-carbon on the equilibrium constants of the dehy- dration reactions shown Fig. 1. Molecular orbital cal- culations show that the inherent substituent effects, as reflected in the gas-phase data, are less drastic. The condensed-phase calculations reproduce the measured values and reveal that the large effects are primarily due to a complex interplay of competing effects con- nected to the solvation process. The calculations pre- dict that the, as yet, unmeasured equilibrium involving (R)-2-hydroxyisocaproyl-CoA very strongly favors the product. Further work is required to determine to what extent this is connected with the successful obser- vation of the proposed penultimate intermediate (5)in the dehydration mechanism (Scheme 1) for reaction A and not for reactions B or C. Experimental procedures Chemical synthesis of CoA thioesters Acrylyl-CoA was synthesized under a gentle stream of N 2 by reacting a three-fold molar excess of acrylyl chloride in dry acetonitrile with free CoASH dissolved in 0.5 mL of aqueous 0.5 m NaHCO 3 . The solution was stirred at room temperature until no yellow color of free thiol was obtained with 5,5¢-dithiobis(2-nitrobenzoate) (Nbs 2 ) [22,23,24]. The pH was adjusted with 5 m HCl to 2 and the solution was stored at )20 °C. In view of the instability of the com- pound, it was prepared and purified the day before use. 2-Hydroxyglutaryl-CoA was synthesized from commer- cial (R)-2-oxo-tetrafuran carboxylic acid that was converted to the corresponding acid chloride by reacting with an excess of oxalyl chloride at 60 °C for 3 h. The excess oxalyl chloride was removed by evaporation under reduced pres- sure. Then, a three-fold excess of the acid chloride was dissolved in dry acetonitrile and reacted with CoASH in 0.5 m NaHCO 3 at room temperature, and the pH was low- ered to 6 to obtain the lactone-CoA. Finally, the lactone- CoA was equilibrated with (R)-2-hydroxyglutaryl-CoA at pH 1 and 25 °C for 3 h. The equilibration was stopped by raising the pH to 8.0. The resulting (R)-2-hydroxyglutaryl- CoA contained about 3% lactone as analyzed by MALDI- TOF MS. 3-Pentynoyl-CoA was synthesized by the procedure used for making 3-pentynoyl pantetheine [15,9]. CoASH (40 lmol) was suspended in 5 mL of dry acetone (CaSO 4 ). Another flask contained 60 lmol of dicyclohexylcarbodii- mide in 5 mL of dry tetrahydrofuran (THF). To the THF solution was added 0.1 g of 3-pentynoic acid. The THF and acetone solutions were mixed quickly and stirred overnight in a sealed flask at 4 °C. The solution was filtered on a sin- tered glass filter to remove dicyclohexylurea. The solvent was removed with a rotary evaporator to yield oil, which was treated with diethyl ether and dried under vacuum. Butyryl-CoA, propionyl-CoA and acetyl-CoA were synthesized from their anhydrides [25]. To CoA (25 lmol) in 1 mL of 0.5 m KHCO 3 ,35lmol of the respective anhy- dride in 0.5 mL of acetonitrile was added. After dilution to 5 mL with water, the mixture was reacted at room tempera- ture until the Nsb 2 test was negative, and then acidified to pH 2. (R)-Lactyl-CoA was synthesized from (R)-lactate and CoASH using 1,1¢-carbonyldiimidazole [17]. Enzymatic synthesis of glutaconyl-CoA with glutaconate CoA-transferase The incubation contained, in 5 mL of 50 mm potassium phosphate (pH 7.0), 20 lmol of acetyl-CoA, 300 lmol of glutaconic acid, and 5 U of recombinant transferase. After 1 h at 37 °C, the mixture was acidified to pH 2 and filtered through a 1 kDa membrane (Amicon; Amersham Equilibrium between hydroxyacyl-CoA and enoyl-CoA A. Parthasarathy et al. 1742 FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS Biosciences, Freiburg, Germany; now part of GE Health- care, Munich, Germany) to remove precipitated protein. Purification of CoA thioesters by reverse-phase chromatography All CoA thioesters were purified by reverse-phase chroma- tography through Sep-Pak C 18 columns (Waters, Milford, MA, USA). The reaction mixtures at pH 2 were freed from solvents under reduced pressure and from precipitated proteins by ultrafiltration. They were then loaded onto C 18 columns washed with methanol and equilibrated with 0.1% (v ⁄ v) trifluoroacetic acid (TFA). After washing with three volumes of the same solution, elution was performed with 0.1% TFA in 50% acetonitrile (v ⁄ v). The eluted CoA esters were freed from acetonitrile on a Speed-Vac concen- trator (Bachofer, Reutlingen, Germany) and vacuum-dried on a lyophilizer (Alpha1-4; Christ Instruments, San Diego, CA, USA). The lyophilized powders were stored at )80 °C until further use. MALDI-TOF MS The CoA thioester samples were purified as described above, and the lyophilized samples were dissolved in 10–40 lL of water. Acetyl-CoA or free CoA was used as internal standard. The matrix was a-cyano-4-hydroxycin- namic acid (Sigma) dissolved in 70% acetonitrile ⁄ 0.1% TFA. One microliter of each sample was mixed with 1 lL of a-cyano-4-hydroxycinnamic acid or a-cyano-3-hydroxy- cinnamic acid as matrix, and spotted onto a gold plate in a dilution series. Measurements were performed with a 355 nm laser in positive reflector mode with a delayed extraction and a positive polarity on the Proteomics Ana- lyzer 4800 mass spectrometer (Applied Biosystems, Fra- mingham, MA, USA) at the MPI for Terrestrial Microbiology, Marburg, Germany. The acceleration voltage was 20 000 V, the grid voltage was 58%, and the delay time was 50 ns. The ratio of reflector voltage was 1.00–1.12. An average of 0.5% of acceleration was laid on the guidewire. The mass range measured was 700–1000 Da. For each spec- trum, more than 1000 shots were accumulated. Enzymatic assays All spectrophotometric assays were performed on Ultro- spec 1100 pro spectrophotometers from Amersham Bio- sciences, installed under aerobic or anaerobic conditions as needed, or a Uvikon 943 double-beam spectrophotometer from Kontron Instruments (Zurich, Switzerland). Quartz cuvettes were used for measurements below 320 nm, and disposable plastic cuvettes for measurements above 320 nm. 2-Hydroxyglutaryl-CoA dehydratase activity was mea- sured under strict anaerobic conditions (d = 1 cm, total volume 0.5 mL at 25 °C) with 50 mm Tris ⁄ HCl (pH 8.0), 5mm MgCl 2 ,5mm dithiothreitol, 0.4 mm ATP, and 0.1 mm dithionite, as well as the activator from A. fermen- tans and dehydratase from C. symbiosum. After incubation for 5 min, the reaction was started by addition of (R)-2-hy- droxyglutaryl-CoA. The formation of (E)-glutaconyl-CoA was measured at 290 nm (e 290 nm = 2.2 mm )1 Æcm )1 ) [7]. In the reverse direction, (E)-glutaconyl-CoA was used as sub- strate. The kinetic constants were determined with 2.0 lgof dehydratase (specific activity of 54 lmol min )1 Æmg )1 pro- tein) and 0.6 lg of activator, using either 0.02–1.0 mm (R)- 2-hydroxyglutaryl-CoA or 0.1–5.0 mm glutaconyl-CoA. Under these conditions, the minimum substrate ⁄ enzyme ratio was 450 : 1. The data were fitted to the Michaelis– Menten equation using the excel program. In the routine assays during purification of the dehydratase, (R)-2-hy- droxyglutaryl-CoA was replaced by (R)-2-hydroxyglutarate, acetyl-CoA, and glutaconate CoA-transferase. Prior to the assay of lactyl-CoA dehydratase from C. pro- pionicum, the crude enzyme fractions or the purified enzyme were incubated for 30 min under anaerobic conditions with 5mm 3-pentynoyl-CoA, which is a reported inactivator of acrylyl-CoA reductase [9], whose activity interferes with the assay under the applied reducing conditions [26]. The pro- tein fraction was freed from the inhibitor by passing it over a 1 mL PD-10 Spintrap G-25 column (GE Healthcare) equilibrated with anaerobic buffer, and concentrating via a Centricon 30 kDa filter (Millipore Corporation, Billerica, MA, USA). The assay was then performed exactly as that for 2-hydroxyglutaryl-CoA dehydratase, except that acrylyl- CoA or lactyl-CoA was used as substrate. The recombinant activator from A. fermentans could be used instead of the activator from C. propionicum, which is very unstable and has never been purified completely [9]. The kinetic constants were determined with 2.0 lg of dehydratase (specific activity of 85 lmolÆmin )1 Æmg )1 protein) and 0.6 lg of activator, using either 0.2–10 mm lactyl-CoA or 0.01–2.0 mm acrylyl- CoA, and evaluated as above. Under these conditions, the minimum substrate ⁄ enzyme ratio was 370 : 1. Acrylyl-CoA reductase activity was measured with propionyl-CoA and ferricenium hexafluorophosphate as electron acceptor [26]. The concentrations of CoASH, acetyl-CoA and glutaco- nyl-CoA (or any other CoA-ester substrate of glutaconate- CoA transferase) were determined in a single assay using Nbs 2 , oxaloacatate, citrate synthase, and transferase [23,24,27]. Similarly, lactyl-CoA and acrylyl-CoA were determined in the same assay, with glutaconate-CoA trans- ferase being replaced by propionate CoA-transferase [28]. Enzyme purification Prior to use, columns, Centricon filters, centrifuge tubes, pipette tips and other plastic materials were stored in a glovebox (Coy Labs, Ann Arbor, MI, USA) for at least A. Parthasarathy et al. Equilibrium between hydroxyacyl-CoA and enoyl-CoA FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS 1743 24 h. All operations during the purifications were per- formed in this box. Buffers and solutions were degassed under reduced pressure, purged with nitrogen, and prere- duced with 2 mm dithiothreitol. Protein was determined by the Bradford method [29]. C. symbiosum HB25 was grown on glutamate, yeast extract, thioglycollate, and biotin [30], whereas C. propioni- cum DSM 1682 required alanine, yeast extract, and cysteine [26]. Production and purification of the recombinant activa- tor from A. fermentans has been described elsewhere [31]. Purification of 2-hydroxyglutaryl-CoA dehydratase from C. symbiosum [32] Frozen cells were suspended in 50 mL of buffer A (50 mm Mops, pH 7.2) under anaerobic conditions and disrupted by ultrasonication in the anaerobic chamber. The cell-free extract was clarified by centrifugation at 100 000 g for 1 h at 4 °C on an Optima L-90K Ultracentrifuge (Beckman Coulter, Brea, CA, USA), and applied to a DEAE–Sepha- rose column equilibrated with buffer A. The column was washed with buffer A, and elution was performed by run- ning a linear gradient of 0–0.7 m NaCl. The active fractions eluted around 0.35 m NaCl. The pooled fractions were combined, and desalted by filtration through a Centricon membrane (30 kDa cut-off); solid ammonium sulfate was then added to 1 m final concentration. This solution was loaded onto a phenyl–Sepharose column pre-equilibrated with buffer A containing 1 m ammonium sulfate. After washing, the proteins were eluted with a gradient of 1– 0.3 m ammonium sulfate. The active fractions starting from 0.5 m ammonium sulfate were combined, desalted, and loaded onto a Q-Sepharose column pre-equilibrated with buffer A. After washing of the column, elution was per- formed with a gradient of 0–0.5 m NaCl. The most active and pure fractions, which eluted around 0.3 m NaCl, were combined, desalted, concentrated, and stored at )80 °C until further use; the yield was 34% (based on the activity of the cell-free extract). Purification of lactyl-CoA dehydratase from C. propionicum The following buffers were used: A, 25 mm Tris ⁄ HCl, 1mm MgCl 2 ,1mm EDTA, and 2 mm dithiothreitol; B, 1.5 m NaCl in A; C, 1.5 m ammonium sulfate in A; and D, 150 mm NaCl in A. Pre-equilibration of the columns allowed purification in about one working day. Frozen cells (12 g) were suspended in 20 mL of buffer A and opened by ultrasonication. The cell-free extract was clarified by ultra- centrifugation for 45 min at 100 000 g and applied onto a Source 15Q column (1.6 · 15 cm) equilibrated with buf- fer A. After washing of the column with 25 mL of buf- fer A, the proteins were eluted in a linear gradient of 0–0.33 m NaCl with 100 mL of buffer B. Two brownish peaks were obtained. The first eluting peak contained the activator (5 mm ATP ⁄ MgCl 2 was added, and the fractions were stored on ice), and the second peak was found to be lactyl-CoA dehydratase. The relevant fractions were com- bined, desalted, and stored on ice. After addition of solid ammonium sulfate to a final concentration of 1.5 m, the solution was sterile-filtered and applied to a Source 15Phe column (1.0 · 10 cm) equilibrated with buffer C. The col- umn was washed with 20 mL of buffer C, and the proteins were eluted in a gradient of 1.5–0 m ammonium sulfate. The brownish fractions were pooled, and the sample was concentrated to about 400 lL with a 100 kDa cut-off Centricon membrane. The concentrated sample (200 lL) was applied to a Superdex 200 column (HR 1.0 ⁄ 30) equili- brated with buffer D, and 0.5 mL fractions were collected. The pure dehydratase (yield 32%) and the still impure acti- vator were not frozen. The dehydratase preparation lost its activity at a rate of 10–15% per day when stored in the glovebox on ice water. Owing to the brownish color of the dehydratase, a ‘blind’ purification could be performed in order to save time and specific activity. The purity of the enzymes was checked by SDS ⁄ PAGE [33]. Computational methods Calculations were carried out with gaussian 03 [34]. The geometry of each of the model systems shown in Fig. 2 was optimized with the B3-LYP ⁄ 6-31G(d) level of theory. Frequency calculations were performed to derive appropriate thermochemical corrections. Improved relative energies, in the gas phase, were obtained using the G3(MP2) model chemistry [35]. The free energies of solvation [21] were obtained using a polarizable continuum model, with Uni- ted Atom Topological Model cavities at the B3-LYP ⁄ 6-311G(d,p) level of theory [36]. The combination of the G3(MP2) gas-phase free energies [DG  ðgÞ , 1 atm reference] with the free energies of solvation [DG Ã s , 1 molÆL )1 (g) fi 1 molÆL )1 (aq) ] [21] and a reference state correction of RT ln( ~ RT) [1 atm (g) fi 1 molÆL )1 (g) ] yields the relative free energies in solution [DG Ã ðaqÞ ], corresponding to a standard state of 1 molÆL )1 . It is these values that were used to determine K conc , which were corrected to K according to Fig. 1. Gaussian archive entries of the gas- phase and solution calculations of the 1a, 2a, 1b, 2b, 1c, 2c and H 2 O can be found in Table S1. Acknowledgements We gratefully acknowledge support (of D. M. Smith) by the Croatian Ministry of Science (project 098- 0982933-2937) and the EC (FP6 contract 043749). Work in Marburg was supported by the Max-Planck Society, Deutsche Forschungsgemeinschaft and the Equilibrium between hydroxyacyl-CoA and enoyl-CoA A. Parthasarathy et al. 1744 FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS Fonds der chemischen Industrie. We thank T. Selmer, Fachhochschule Aachen, Germany, for advice on the purification of lactyl-CoA dehydratase. References 1 Kim J, Hetzel M, Boiangiu CD & Buckel W (2004) Dehydration of (R)-2-hydroxyacyl-CoA to enoyl-CoA in the fermentation of alpha-amino acids by anaerobic bacteria. FEMS Microbiol Rev 28, 455–468. 2 Buckel W, Hetzel M & Kim J (2004) ATP-driven electron transfer in enzymatic radical reactions. Curr Opin Chem Biol 8, 462–467. 3 Buckel W & Keese R (1995) One electron redox reac- tions of CoASH esters in anaerobic bacteria. A mecha- nistic proposal. Angew Chem Int Ed Engl 34, 1502– 1506. 4 Kim J, Darley DJ, Buckel W & Pierik AJ (2008) An allylic ketyl radical intermediate in clostridial amino- acid fermentation. Nature 452, 239–242. 5 Smith DM, Buckel W & Zipse H (2003) Deprotonation of enoxy radicals: theoretical validation of a 50-year-old mechanistic proposal. Angew Chem Int Ed Engl 42, 1867–1870. 6 Scherf U & Buckel W (1993) Purification and properties of an iron–sulfur and FAD-containing 4-hydroxybuty- ryl-CoA dehydratase ⁄ vinylacetyl-CoA delta 3–delta 2-isomerase from Clostridium aminobutyricum. Eur J Biochem 215, 421–429. 7 Kim J, Darley DJ & Buckel W (2005) 2-Hydroxyisocap- royl-CoA dehydratase and its activator from Clostrid- ium difficile. FEBS J 272 , 550–561. 8 Schweiger G & Buckel W (1985) Identification of acry- late, the product of the dehydration of (R)-lactate catal- ysed by cell-free extracts from Clostridium propionicum. FEBS Lett 185, 253–256. 9 Kuchta RD & Abeles RH (1985) Lactate reduction in Clostridium propionicum. Purification and properties of lactyl-CoA dehydratase. J Biol Chem 260, 13181–13189. 10 Buckel W (1980) The reversible dehydration of (R)-2-hydroxyglutarate to (E)-glutaconate. Eur J Biochem 106, 439–447. 11 Klees AG & Buckel W (1991) Synthesis and properties of (R)-2-hydroxyglutaryl-1-CoA. (R)-2-hydroxyglutaryl- 5-CoA, an erroneous product of glutaconate CoA-trans- ferase. Biol Chem Hoppe Seyler 372, 319–324. 12 Jacob U, Mack M, Clausen T, Huber R, Buckel W & Messerschmidt A (1997) Glutaconate CoA-transferase from Acidaminococcus fermentans: the crystal structure reveals homology with other CoA-transferases. Structure, 5, 415–426. 13 Hans M, Sievers J, Mu ¨ ller U, Bill E, Vorholt JA, Linder D & Buckel W (1999) Cloning and sequencing of the genes of 2-hydoxyglutaryl-CoA dehydratase from Acidaminococcus fermentans. Eur J Biochem 265, 404– 414. 14 Hofmeister AE & Buckel W (1992) (R)-lactyl-CoA de- hydratase from Clostridium propionicum. Stereochemis- try of the dehydration of (R)-2-hydroxybutyryl-CoA to crotonyl-CoA. Eur J Biochem 206, 547–552. 15 Gomes B, Fendrich G & Abeles RH (1981) Mechanism of action of glutaryl-CoA and butyryl-CoA dehydro- genases. Purification of glutaryl-CoA dehydrogenase. Biochemistry 20, 1481–1490. 16 Jones ERH, Mansfield GH & Whiting MC (1954) Researches on acetylenic compounds. Part XLVI. Prototropic rearrangements of acetylenic acids. J Chem Soc 3201–3208. 17 Kawaguchi A, Yoshimura T & Okuda S (1981) A new method for the preparation of acyl-CoA thioesters. J Biochem (Tokyo) 89 , 337–339. 18 Friedrich P, Darley DJ, Golding BT & Buckel W (2008) The complete stereochemistry of the enzymatic dehydration of 4-hydroxybutyryl coenzyme A to croto- nyl coenzyme A. Angew Chem Int Ed Engl 47, 3254– 3257. 19 Sato K, Nishina Y, Setoyama C, Miura R & Shiga H (1999) Unusually high standard redox potential of acrylyl-CoA ⁄ propionyl-CoA couple among enoyl- CoA ⁄ acyl-CoA couples: a reason for the distinct meta- bolic pathway of propionyl-CoA from longer acyl- CoAs. J Biochem (Tokyo) 126, 668–675. 20 Tareke E, Rydberg P, Karlsson P, Eriksson S & To ¨ rnq- vist M (2000) Acrylamide: a cooking carcinogen? Chem Res Toxicol 13, 517–522. 21 Ben-Naim A & Marcus Y (1984) Solvation thermody- namics of nonionic solutes. J Chem Phys 81, 2016– 2027. 22 Ellman GL (1959) Tissue sulfhydryl groups. Arch Bio- chem Biophys 82, 70–77. 23 Riddles PW, Blakeley RL & Zerner B (1979) Ellman’s reagent: 5,5¢-dithiobis-(2-nitrobenzoic acid) – a reexami- nation. Anal Biochem 94, 75–81. 24 Riddles PW, Blakeley RL & Zerner B (1983) Reas- sessment of Ellman’s reagent. Methods Enzymol 91, 49–60. 25 Simon EJ & Shemin D (1953) The preparation of S-suc- cinyl coenzyme A. J Am Chem Soc 75, 2520. 26 Hetzel M, Brock M, Selmer T, Pierik AJ, Golding BT & Buckel W (2003) Acryloyl-CoA reductase from Clostridium propionicum. An enzyme complex of propionyl-CoA dehydrogenase and electron-transferring flavoprotein. Eur J Biochem 270, 902–910. 27 Buckel W, Dorn U & Semmler R (1981) Glutaconate CoA-transferase from Acidaminococcus fermentans. Eur J Biochem 118, 315–321. 28 Selmer T, Willanzheimer A & Hetzel M (2002) Propio- nate CoA-transferase from Clostridium propionicum. A. Parthasarathy et al. Equilibrium between hydroxyacyl-CoA and enoyl-CoA FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS 1745 Cloning of the gene and identification of glutamate 324 at the active site. Eur J Biochem 269, 372–380. 29 Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein uti- lizing the principle of protein-dye binding. Anal Biochem 72, 248–254. 30 Buckel W (1986) Biotin-dependent decarboxylases as bacterial sodium pumps: purification and reconstitution of glutaconyl-CoA decarboxylase from Acidaminococ- cus fermentans. Methods Enzymol 125, 547–558. 31 Hans M, Buckel W & Bill E (2000) The iron–sulfur clusters of 2-hydroxyglutaryl-CoA dehydratase from Acidaminococcus fermentans. Spectroscopic and bio- chemical investigations. Eur J Biochem 267, 7082–7093. 32 Hetzel M (2004) Towards the Mechanism of the 2-hydroxyglutaryl-CoA-dehydratase from Clostridium Symbiosum. PhD thesis, Philipps-University, Marburg, Germany. 33 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680–685. 34 Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Montgomery JA Jr, Vreven T, Kudin KN, Burant JC et al. (2004) Gaussian 03, Revision E.01. Gaussian, Inc., Wallingford, CT. 35 Curtiss LA, Redfern PC, Raghavachari K, Rassolov V & Pople JA (1999) Gaussian-3 theory using reduced Moller–Plesset order. J Chem Phys 110, 4703–4709. 36 Barone V, Cossi M & Tomasi J (1997) A new definition of cavities for the computation of solvation free energies by the polarizable continuum model. J Chem Phys 107, 3210–3221. Supporting information The following supplementary material is available: Table S1. Gaussian archive entries of the gas-phase and solution calculations of the species 1a, 2a, 1b, 2b, 1c, 2c, and H 2 O. This supplementary material can be found in the online version of this article. Please note: As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Equilibrium between hydroxyacyl-CoA and enoyl-CoA A. Parthasarathy et al. 1746 FEBS Journal 277 (2010) 1738–1746 ª 2010 The Authors Journal compilation ª 2010 FEBS . solvated more strongly than the enoyl one (2b), the difference between them, and hence the associated effect on DG Ã ðaqÞ (B), is much smaller than for reaction C. The favorable contribu- tion from DG Ã s (H 2 O). kJÆmol )1 and K a = 1610. Given the good agreement between theory and experiment obtained for equilibria B and C, we are confident that these values provide a reason- ably accurate description of the thermodynamics. of 2-hydroxyisocaproyl-CoA dehydration. They are cer- tainly consistent with the fact that the equilibrium concentration of 1a was not detectable in experiments concerning its conversion into 2a [7]. The precise experimental