- Email: [email protected]
C295A mutant secondary alcohol dehydrogenase from Thermoanaerobacter ethanolicus has broadened substrate specificity for aryl ketones
Archives of Biochemistry and Biophysics 606 (2016) 151e156
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi
I86A/C295A mutant secondary alcohol dehydrogenase from Thermoanaerobacter ethanolicus has broadened substrate specificity for aryl ketones Christopher M. Nealon a, 1, Travis P. Welsh a, 2, Chang Sup Kim b, Robert S. Phillips a, c, * a b c
Department of Chemistry, University of Georgia, Athens, GA 30602, USA Department of Chemical and Biological Engineering, Hanbat National University, Daejeon 34158, South Korea Department of Biochemistry and Molecular Biology, University of Georgia, Athens, GA 30602, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 June 2016 Received in revised form 29 July 2016 Accepted 1 August 2016 Available online 3 August 2016
Thermoanaerobacter ethanolicus secondary alcohol dehydrogenase (SADH) reduces aliphatic ketones according to Prelog's Rule, with binding pockets for small and large substituents. It was shown previously that the I86A mutant SADH reduces acetophenone, which is not a substrate of wild-type SADH, to give the anti-Prelog R-product (Musa, M. M.; Lott, N.; Laivenieks, M.; Watanabe, L.; Vieille, C.; Phillips, R. S. ChemCatChem 2009, 1, 89e93.). However, I86A SADH did not reduce aryl ketones with substituents larger than fluorine. We have now expanded the small pocket of the active site of I86A SADH by mutation of Cys-295 to alanine to allow reaction of substituted acetophenones. As predicted, the double mutant I86A/ C295A SADH has broadened substrate specificity for meta-substituted, but not para-substituted, acetophenones. However, the increase of the substrate specificity of I86A/C295A SADH is accompanied by a decrease in the kcat/Km values of acetophenones, possibly due to the substrates fitting loosely inside the more open active site. Nevertheless, I86A/C295A SADH gives high conversions and very high enantiomeric excess of the anti-Prelog R-alcohols from the tested substrates. © 2016 Elsevier Inc. All rights reserved.
Keywords: Mutagenesis Substrate specificity Stereospecificity Alcohol dehydrogenase Thermophilic
1. Introduction Two alcohol dehydrogenases, one with a specificity for primary alcohols, and a secondary alcohol dehydrogenase, were purified from Thermoanaerobacter ethanolicus, which was isolated from hot springs in Yellowstone National Park [1,2]. The wild-type secondary alcohol dehydrogenase (SADH) follows Prelog's rule in the reduction of ketones, adding the R-hydride of the NADPH to the re-face of the ketone [3]. The substrate specificity of wild-type SADH is restricted to primarily linear aliphatic and alicyclic ketones [4e9]. Fig. 1 shows the active site of SADH from Thermoanaerobacter brockii (formerly Thermoanaerobium brockii [10]), which is identical
Abbreviations: ADH, alcohol dehydrogenase; NADPþ, nicotinamide adenine dinucleotide phosphate; SADH, Thermoanaerobacter ethanolicus secondary ADH. * Corresponding author. Department of Chemistry, University of Georgia, Athens, GA 30602, USA E-mail address: [email protected] (R.S. Phillips). 1 Present address: Department of Chemistry, United States Naval Academy, Annapolis, MD 21402, USA. 2 Present address: Medical College of Georgia, Augusta University, Augusta, GA 30912, USA. http://dx.doi.org/10.1016/j.abb.2016.08.002 0003-9861/© 2016 Elsevier Inc. All rights reserved.
with the enzyme from T. ethanolicus [11]. Thus, a number of active site residues have been mutated in order to broaden the substrate specificity of SADH and make it useful in biocatalytic applications. These mutant SADHs, including S39T, C295A, W110A, and I86A, were found to have altered substrate specificity and stereoselectivity [9,12e14]. The I86A mutation of SADH not only expands substrate specificity to include acetophenone, which is a very poor substrate for wild-type SADH, but also reverses the usual preferred stereochemistry to produce the anti-Prelog R-product [14]. Unfortunately, I86A SADH exhibits limited reactivity with substituted acetophenones, fluorine being the only tolerable substituent found in the initial study. Homochiral 1-arylalkanols are useful intermediates in the preparation of pharmaceuticals. Thus, it was of interest to expand the active site further to allow the reduction of substituted acetophenones. The C295A mutant SADH was studied previously by Heiss and coworkers, and was found to increase the size of the alkyl group which can bind in the “small pocket” by one carbon atom [12]. Due to the proximity of Cys-295 to Ile-86 (Fig. 1), we predicted that having both mutations in the active site would expand the size of the small pocket to allow binding and reaction of
152
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156
Fig. 1. Stereoview of TbSADH, with residues of interest labeled. NADPþ shown in stick-form and zinc as a cyan sphere. This image was prepared with Pymol (The PyMOL Molecular €dinger, LLC) using the PDB file (1YKF). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version Graphics System, Version 1.7 Schro of this article.)
ring-substituted acetophenones with subsitutents larger than fluorine. To test this hypothesis, we have now prepared the I86A/ C295A double mutant of SADH, and we have found, as expected, that it has broadened specificity for substituted acetophenones. 2. Materials and methods
isolated from the host E. coli DH5a cells. Using the Quikchange method, the pADHB25-I86A plasmid was further mutated with the following forward and reverse primers, with the mutation indicated in bold. The resulting plasmid was sequenced to confirm the presence of the C295A mutation. C295A-F (ATAAAAGGCGGGCTAGCCCCCGGTGGACG). C295A-R (TTTCCGCCCGATCGGGGGCCACCTGCAGA).
2.1. General methods 2.4. Purification of secondary alcohol dehydrogenase Gas chromatography was performed with a Varian 3300 GC (Agilent Tech; Santa Clara, CA) using a Supelco (Sigma Aldrich; St. Louis, MO) b-Dex 120 cyclodextrin chiral column (30 m, 0.25 mm [i.d.], 0.25 mm film thickness) with He as the carrier gas and equipped with a flame ionization detector. Kinetic experiments and assays were performed on a Varian Cary 100 UVevisible spectrophotometer (Agilent Tech; Santa Clara, CA) equipped with a Peltier thermoelectric temperature-controlled 12-cell holder. 1H and 13C NMR analyses were collected on a Varian 400 MHz spectrometer with CDCl3 as the solvent at room temperature with tetramethylsilane or the solvent peak as the reference. 2.2. Materials Substrates were used as purchased from commercial suppliers with the exception of 3’-bromoacetophenone, which was prepared by a published procedure [15]. Acetophenone was purchased from Fisher Scientific (Waltham, MA). 20 ,40 -Difluoroacetophenone was bought from Acros (Geel, Belgium). The 2-acetylpyridine was purchased from Pfaltz and Bauer (waterbury, CT). The rest of the substrates tested were purchased from Sigma-Aldrich (St. Louis, MO). Acetonitrile was a product of Burdick and Jackson (Morristown, NJ). NADPH was purchased from Acros (Geel, Belgium). The antibiotics, kanamycin and ampicillin, were bought from Roche (Indianapolis, IN) and Fisher Biotech (fair lawn, NJ), respectively. Commercial grade solvents were used without further purification. The RedAagarose was obtained from Sigma-Aldrich (St. Louis, MO). 2.3. Mutagenesis The I86A plasmid, made as reported previously [14], was
The protein was expressed and purified based on our previously published method [9]. Modifications made to this method include the use of 50 mM Tris HCl, pH 8.0, buffer containing 5 mM DTT and 10 mM ZnCl2, a 5 mL RedA-agarose column, and optimization of the low salt (0.02 M NaClO4) and high salt (0.2 M NaClO4) wash and elution solutions. No further purification was needed after the RedA-agarose column as polyacrylamide gel electrophoresis showed a single band. The pooled fractions were stored at 80 C. 2.5. Enzyme assays The enzyme was assayed as previously described [14]. The enzyme assays were performed at 50 C in triplicate. The enzyme activity was measured in 50 mM potassium phosphate buffer (pH 6.5 at 50 C) with 0.4 mM NADPH to follow ketone reduction. The initial velocity was recorded on a Varian Cary 100 UV/Vis spectrophotometer at 340 nm for 10 min by monitoring NADPH consumption (ketone reduction). The enzyme was preincubated in 50 mM buffer solution with 5 mM DTT and 10 mM ZnCl2 at 50 C in the UV/Vis sample compartment for 10 min before addition to the assay mixtures. The substrate stock solutions were prepared in acetonitrile. Each assay contained not more than 5% acetonitrile, since higher concentrations were found to inhibit the enzyme activity. The kinetic data were fit to the Michaelis-Menten equation (Equation (1)) with the HYPERO program of Cleland [16]. The enzyme concentrations were determined from the A280 value of 0.82 for a 0.1% solution, and using the subunit molecular weight of 37.7 kDa.
v ¼ Vmax ½S=ðKm þ ½SÞ
(1)
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156
one carbon to propiophenone results in only a 2-fold decrease in kcat/Km. However, with increasing size of substituents on the ring, the kcat/Km value decreases dramatically. Interestingly, we found in this study that 3’-chloroacetophenone, which we had not tested previously, reacts with I86A mutant SADH, although the kcat/Km value is decreased by about 36-fold compared to acetophenone (Table 1). We had found previously that acetophenones with other substituents larger than fluorine are not detectable as substrates for I86A SADH, which was confirmed in this study (Table 1). Another interesting result is that 2-acetylpyridine, 4-acetylpyridine and 3acetylthiophene react much better than 3-acetylpyridine and 2acetylthiophene with I86A SADH. Due to the similar overall size of these heterocyclic substrates, these differences must be due to electronic effects rather than steric effects. The electronwithdrawing effects of the pyridine ring on the carbonyl are strongest in 2- and 4-acetylpyridine, due to resonance contributions, while the electron-donating effect of thiophene is weakest for 3-acetylthiophene. These observations suggest that hydride transfer to the carbonyl is at least partially rate-determining for these substrates and facilitated by making the carbonyl carbon more electron-deficient. The aromatic ring of these substrates must bind in the “small pocket” of I86A SADH, since the anti-Prelog R-enantiomer is the observed reduction product. We incorporated the C295A mutation into the I86A mutant SADH to expand the small pocket of the active site even further, and thus we hoped to broaden the scope of acetophenone substrates that can be reduced. While the I86A/C295A double mutant SADH unexpectedly showed no detectable reduction with para-substituted acetophenones, meta-substitution was accepted, as shown in Table 1. Small ortho-substituents are also tolerated, since 2’-methylacetophenone was found to be a slow substrate (Table 1). It is interesting that the kcat/Km value for unsubstituted acetophenone is reduced about 8-fold for I86A/ C295A SADH compared to I86A SADH. This suggests that the decrease in substrate specificity is gained at the expense of reaction efficiency. In contrast, 3’-chloroacetophenone is about 3-fold better as a substrate for I86A/C295A SADH than for I86A SADH. The Km values for all acetophenones with I86A/C295A SADH were
2.6. Asymmetric reduction with I86A and I86A/C295A Reactions included in 1 mL, substrate (0.21 mmol), 1 mg NADPH, 0.5 mL Tris-HCl buffer (8.0 pH with 10 mM ZnCl2), 0.2 mL 2propanol, 1 mg DTT and I86A or I86A/C295A SADH (0.654 mg). The reactions were incubated for 24 h at 50 C in a shaker. After 24 h, the reactions were quenched by addition of dichloromethane. This organic extract could be used to quantify the percent conversions by gas chromatography. The alcohol products were identified by comparison with the standards obtained from NaBH4 reduction of the ketones. Acetylation of the reaction products with acetic anhydride and pyridine was performed [17]. The enantiomeric excess of the products was found by GC analysis after acetylation. The ee values were calculated from the peak areas of the R and Sacetates, using Equation (2).
ðAreaðRÞ AreaðSÞÞ=ðAreaðRÞ þ AreaðSÞÞ* 100
153
(2)
2.7. Hammett-Taft calculations The curve fitting tool (cftool) in MATLAB (Version R2014b) was used to fit the three sets of data (Hammett, Taft, and either log(kcat/ (kcat)0) or log((kcat/KM)/(kcat/KM)0)) to the equation z ¼ a*x þ b*y. Equation (3) fitting parameters: 1.013 (SSE), 0.1435 (R2), 0.02113 (Adj. R2), 0.3805 (RMSE). Equation (4) fitting parameters: 0.2519 (SSE), 0.6082 (R2), 0.5522 (Adj. R2), 0.1897 (RMSE). Equation (3)a ¼ 1.531 (2.595, 5.656) b ¼ 1.312 (0.3748, 2.25). Equation (4)a ¼ 1.52 (0.5461, 3.586) b ¼ 1.105 (0.6524, 1.558). 3. Results and discussion In our previous report on the substrate specificity of I86A SADH, we did not evaluate kinetic parameters for all of the aromatic substrates [14]. As shown in Table 1, I86A SADH has the highest specificity constant, kcat/Km, with unsubstituted acetophenone, consistent with our previous results. Extension of the alkyl chain by
Table 1 I86A and I86A/C295A SADH kinetic parameters. O R1
R2 NADPH
Substrate
OH
SADH R1
R2
NADP+
I86A/C295A
I86A
R1
R2
kcat (s1)
kcat (s1) (M1s1)
kcat (s1)
kcat/Km (M1s1)
CH3 C6H5 30 -ClC6H5 30 -BrC6H5 30 -CH3C6H5 30 -IC6H5 20 -CH3C6H5 2-pyridyl 3-pyridyl 4-pyridyl 2-thienyl 3-thienyl C6H5 2,4-F2C6H3 30 -CH3OC6H5 30 -CF3C6H5
CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH2CH3 CH3 CH3 CH3
35 ± 6 15.8 ± 1.6 12.3 ± 2.4 6.5 ± 1.1 2.0 ± 0.33 0.61 ± 0.06 0.18 ± 0.04 2.31 ± 0.23 N/Aa 13.1 ± 1.9 1.5 ± 0.2 N/Aa 23.3 ± 5.5 2.9 ± 0.7 0.08 ± 0.02 0.04 ± 0.008
1080 ± 110 1900 ± 220 1300 ± 250 650 ± 110 180 ± 28 160 ± 30 18 ± 3 260 ± 30 630 ± 14 2400 ± 400 260 ± 60 350 ± 60 1500 ± 150 230 ± 60 210 ± 40 25 ± 10
1.01 ± 0.16 8.2 ± 0.6 0.66 ± 0.09 <0.01 <0.01 <0.01 <0.01 6.9 ± 1.6 0.47 ± 0.15 1.7 ± 0.1 1.7 ± 0.2 14.9 ± 4.2 6.5 ± 5.5 2.2 ± 0.7 <0.01 <0.01
1330 ± 340 14900 ± 3100 410 ± 90 <1 <1 <1 <1 1600 ± 260 200 ± 55 1050 ± 101 510 ± 85 4300 ± 730 8100 ± 1700 620 ± 140 <1 <1
a
Substrate inhibition precluded determination of kcat.
154
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156
generally around 8e10 mM, with some exceptions, so the kcat value is most affected by substitution, as can be seen in Table 1. The reactivity of the acetophenones also decreases with increasing size of the meta-substituent, and the meta-substituent can be as large as iodo, methoxy or trifluoromethyl. Since it appeared that both electronic and steric effects influence the reactivity of acetophenone substrates, a dual-parameter Hammett-Taft analysis [18] was performed with the kcat and kcat/ Km data for I86A/C295A SADH in Table 1. The plots of the results are shown in Fig. 2. The three-dimensional fit to the general equation z ¼ ax þ by for each was as follows:
Log Kcat ðKcat Þ0 ¼ 1:531s þ 1:312Es
(3)
Log ðKcat =Km Þ Kcat ðKcat Þ0 ¼ 1:531s þ 1:312Es
(4)
The sensitivity factors r and d (a and b of the general equation above) define how sensitive the kinetic parameters are to Hammett s and Taft ES, respectively. From each of the best fit planes, there is similar impact from the electronic effects as well as the steric effects on the reaction. The value of r is identical for both kcat and kcat/ Km, 1.53 and 1.52, respectively, in Equations (3) and (4), showing that both kinetic parameters are affected similarly by electronic effects. This suggests that the hydride transfer step for acetophenone reduction is partially rate-determining for both kcat and kcat/ Km. The positive value of r indicates that electron-withdrawal increases substrate reactivity, consistent with the results from
acetylpyridines
and
acetylthiophenes
discussed
above.
The
d values, 1.312 and 1.105, are slightly less than the r values and not significantly different. This suggests that the larger substituents may not allow the substrate to adopt the optimal geometry for hydride transfer on binding to the active site. Another comparison of the efficiency of the two mutant enzymes is through preparative reactions, as we had done previously. Table 2 compares the previously published I86A SADH data [14], with two additional substrates, and the new I86A/C295A SADH data. These reactions contained 0.21 mmol of substrate, corresponding to 25.2 mg in the case of acetophenone. In general, the preparative reaction results correlate well with the kinetic results in Table 1. As expected, 3’-chloroacetophenone reacts much better with I86A/C295A SADH than with I86A SADH, giving conversions of 23% and 0.6%, respectively. 2-Acetylpyridine and 4-acetylpyridine gave fairly high conversion with I86A/C295A SADH, although slightly lower conversion than the previously published results with I86A [14]. One interesting observation was that the percent conversion with 2-acetylthiophene and 3-acetylthiophene rose with 2-acetylthiophene with I86A/C295A SADH, while the percent conversion fell with 3-acetylthiophene with I86A/C295A SADH. 30 Aminoacetophenone, 3’-nitroacetophenone, 3’-ethoxyacetophenone, and 4’-methylacetophenone did not show any detectable reduction product with I86A/C295A SADH, and thus would be too large for the active site of I86A SADH. In all cases where reaction was observed, we found the product to be the antiPrelog R-alcohol with >99% ee.
Fig. 2. I86A/C295A SADH Hammett-Taft Plot. Top: Plot of substitutent effects on kcat. Bottom: Plot of substituent effects of kcat/Km.
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156 Table 2 I86A and I86A/C295 SADH GC assays. O
OH
SADH R2
R1
R2
R1
NADPH
NADP+
O
OH SADH
Substrate
I86A/C295A SADH
I86A SADH
R1
R2
Conv. (%)
ee (%)
R/S
Conv. (%)
ee (%)
R/S
C6H5 30 -ClC6H5 30 -BrC6H5 30 -IC6H5 30 -CH3C6H5 20 -CH3C6H5 30 -CH3OC6H5 30 -CF3C6H5 2-pyridyl 3-pyridyl 4-pyridyl 2-thienyl 3-thienyl C6H5 2,4-F2C6H3
CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH2CH3 CH3
68 23 8 4 23 <1 5 3 53 75 98 27 39 31 4
>99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99
R R R R R R R R R R R R R R R
47a 0.6 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 >99a 46a >99a 2.52 76a 60a 33a
98a >99 e e e e e e >99a >99a >99a >99 >99a >99a >99a
Ra R e e e e e e Ra Ra Ra R Ra Ra Ra
a
From Musa et al. [14].
Reetz et al. have also performed mutagenesis studies with SADH from T. brockii in order to change the substrate specificity and stereoselectivity [19]. For the reduction of axially chiral cyclohexanones, which are reduced by wild-type SADH with good yield but modest enantioselectivity (66% ee), they found that the W110T mutant SADH produced the R-alcohols with high conversions and >92% ee. Conversely, the I86A mutant SADH discussed in this paper provided the S-alcohols in some cases, with stereoselectivities ranging from modest to excellent. Recently, they used triple code saturation mutagenesis to optimize the reduction of small saturated heterocyclic ketones which are difficult to reduce [20]. The residues in the active site targeted for mutagenesis included Ala-85, Ile-86, Trp-110, Leu-294, and Cys-295. After an initial round of saturation mutagenesis of each of these sites, the best mutations were grouped into two sites for a second round of mutagenesis. Site A (A85/I86/L294/C295) was mutated using a V-N-L triple code, and site B (A85/I86/W110/L294) was mutated with a V-Q-L triple code. A double mutant, I86N/C295N, was identified from the site A library that reduces the ketones to the R-alcohols in with very high ee. A triple mutant, I86V/W110L/L294Q, was found in the site B library that provide the S-alcohols in 95% ee. Mutagenesis has been used to modify substrate specificity of other dehydrogenases. The carbonyl reductase from Candida parapsilosis (CPCR2) contains a zinc ion, as does SADH, and also has
155
activity with medium chain alcohols. Jakoblinnert and coworkers expanded the active site of CPCR2, and thus affected the substrate specificity [21]. Five residues were selected for site-saturation mutagenesis (Leu-55, Pro-92, Gly-118, Leu-119, Leu-262). Of the active mutants at those locations, only L119 M gave activity that was greater than wild-type CPCR2. Even though methionine and leucine have side chains of similar size, the lack of branching in the methionine side chain provided a less constraining CPCR2 active site. Zhu et al. obtained a crystal structure of the carbonyl reductase, and using the structure as a guide, prepared three mutants, F285A, W286A, and F285A/W286A [22]. All of these mutant enzymes showed a reversal of stereochemistry in the reduction of aryl ketones, giving the anti-Prelog products. Zhu and coworkers also studied Sporobolomyces salmonicolor carbonyl reductase (SsCR), with the goal to improve the activity of wild-type SsCR towards para-substituted acetophenones [23]. They focused their site-saturation mutagenesis on the Gln-245 residue as a result of docking studies. The mutant enzymes with the greatest improvement were Q245H, Q245P, and Q245L, all three of which interestingly switched the stereochemistry of the alcohol product from R to S. They utilized site-saturation mutagenesis on the Met-242 residue to probe the SsCR active site further, and found improved properties for some of the substrates tested [23]. Also, M242Y, M242D, M242C and M242G mutations reversed the stereochemistry of the alcohol product. Since both Met-242 and Gln245 mutations alter the substrate specificity separately, they opted to design a double mutant at those sites. M242L/Q245P and M242L/ Q245T showed the largest boost in ee [23]. Broadening the substrate scope to include benzophenones, Li and coworkers found that mutagenesis at the Gln-245 residue improved substrate specificity as well as providing the S-isomer [24]. Zhang and coworkers recently reported a mutant SsCR study with b-aminoketones [25]. While there was a little activity with M242F/Q245T and b-aminoacetophenone, a docking study followed by sitesaturation mutagenesis yielded useful mutations at Pro-170 and Leu-174. L174Y and L174W were the single mutations with the greatest activity against b-aminoacetophenone. In addition, they designed double mutant enzymes of Pro-170 and Leu-174, with the most interesting mutant SsCRs being P170R/L174Y, P170R/L174W, and P170H/L174W. When each of these mutated enzymes were tested against b-amino-2-acetylthiophene, only L174Y and P170R/ L174Y gave high ee for each of the two substrates [25]. Kavanagh and coworkers studied the active site of Candida tenuis xylose reductase (CtXR) [26], which was later found to react with a-ketoesters [27]. Through modeling studies, the Trp-23 appeared to conflict with the a-carbonyl of the substrates. The tryptophan was replaced by phenylalanine and then tyrosine using site-directed mutagenesis. The activity was increased for each of these mutants when compared with wild-type CtXR against the various a-ketoesters [27]. Kratzer and researchers studied the kinetics of acetophenone and other assorted ketones reduction with W23F, W23Y and wild type CtXR [28]. While the mutations decreased the activity with most of the ketones, oxopantoyl lactone
Fig. 3. Cartoon showing the binding of meta-substituted acetophenone in the small pocket of I86A/C295A SADH.
156
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156
and ethyl benzoylformate interestingly increased in activity of four to five-fold and two to three-fold, respectively. Due to the mixed kinetic results, the researchers rationalized that the Trp-23 has a role in selective binding and the specificity of the enzyme [28]. 4. Conclusions Taken together, our present results and the previously published results demonstrate that rational mutagenesis and directed evolution are practical approaches to broaden substrate specificity of alcohol dehydrogenases for reduction of a wide range of nonphysiological substrates. The wild-type SADH active site cannot accommodate a phenyl moiety in the small pocket, since wild-type SADH shows negligible activity with acetophenone in our assays [14]. Thus, the observation that only a single mutation (I86A) could open the small pocket enough for a phenyl ring to fit inside and give a substrate with a kcat/Km value of >104 M1 s1 (Table 1) is remarkable. The “small pocket” of wild-type SADH allows binding of substituents only as large as propyl or isopropyl [4,6,12]. The I86A mutation removes three carbons from the small pocket, and now the substrate of I86A SADH can have a six-carbon phenyl substituent. Hence, one can consider the active site of these enzymes essentially as a rigid scaffold, the shape of which can be tailored by mutagenesis to accommodate a wide range of substrates. We have now shown that incorporation of an additional mutation (C295A) of a neighboring residue in the I86A SADH active site has expanded the pocket to allow reaction of m-substituted aryl ketones (Fig. 3). The substituents that are tolerated are sterically similar to or smaller than the sulfur atom that was removed by mutagenesis. It is noteworthy that in increasing the size of the active site through multiple mutations, the efficiency of the enzyme decreased even as the substrate pool was able to increase. This could be because the substrate has more freedom of motion within the expanded active site. In conclusion, this I86A/C295A double mutation provides a starting point for further expansion of the SADH active site to allow reaction of both m- and p-substituted aromatic ketones.
Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.abb.2016.08.002. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
[22] [23] [24] [25]
Acknowledgements [26]
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
[27] [28]
J.G. Zeikus, P.W. Hegge, M.A. Anderson, Arch. Microbiol. 122 (1979) 41e48. J. Wiegel, L.G. Ljungdahl, Arch. Microbiol. 128 (1981) 343e348. V. Prelog, Pure Appl. Chem. 9 (1964) 119e130. E. Keinan, E.K. Hafeli, K.K. Seth, R. Lamed, J. Am. Chem. Soc. 108 (1986) 162e169. V.T. Pham, R.S. Phillips, L.G. Ljungdahl, J. Am. Chem. Soc. 111 (1989) 1935e1936. V.T. Pham, R.S. Phillips, J. Am. Chem. Soc. 112 (1990) 3629e3632. C.S. Zheng, V.T. Pham, R.S. Phillips, Bio. Med. Chem. Letts. 2 (1992) 619e622. C. Zheng, V.T. Pham, R.S. Phillips, Catal. Today 22 (1994) 607e620. A.E. Tripp, D.S. Burdette, J.G. Zeikus, R.S. Phillips, J. Am. Chem. Soc. 120 (1998) 5137e5141. Y.E. Lee, M.K. Jain, C. Lee, J.G. Zeikus, Int. J. Syst. Evol. Microbiol. 43 (1993) 41e51. C.M. Nealon, M.M. Musa, J.M. Patel, R.S. Phillips, ACS Catal. 5 (2015) 2100e2114. C. Heiss, M. Laivenieks, J.G. Zeikus, R.S. Phillips, Bioorg. Med. Chem. 9 (2001) 1659e1666. M.M. Musa, K.I. Ziegelmann-Fjeld, C. Vieille, J.G. Zeikus, R.S. Phillips, J. Org. Chem. 72 (2007) 30e34. M.M. Musa, N. Lott, M. Laivenieks, L. Watanabe, C. Vieille, R.S. Phillips, ChemCatChem 1 (2009) 89e93. D.E. Pearson, H.W. Pope, W.W. Hargrove, Org. Syn. 40 (1960) 7. A. Ghanem, V. Schuring, Tetrahedron Asymmetry 14 (2003) 57e62. W.W. Cleland, Methods Enzymol. 63 (1979) 103e138. R.W. Taft Jr., In “Steric Effects in Organic Chemistry”, in: M.S. Newman (Ed.), John Wiley & Sons, Inc., New York, N. Y.,, 1965, p. 565. R. Agudo, G.D. Roiban, M.T. Reetz, J. Am. Chem. Soc. 135 (2012) 1665e1668. Z. Sun, R. Lonsdale, A. Ilie, G. Li, J. Zhou, M.T. Reetz, ACS Catal. 6 (2016) 1598e1605. A. Jakoblinnert, J. Wachtmeister, L. Schukur, A.V. Shivange, M. Bocola, M.B. Ansorge-Schumacher, U. Schwaneberg, Protein Eng. Des. Sel. 26 (2013) 291e298. D. Zhu, Y. Yang, S. Majkowicz, T.H. Pan, K. Kantardjieff, L. Hua, Org. Lett. 10 (2008) 525e528. H. Li, Y. Yang, D. Zhu, L. Hua, K. Kantardjieff, J. Org. Chem. 75 (2010) 7559e7564. H. Li, D. Zhu, L. Hua, E.R. Biehl, Adv. Synth. Catal. 351 (2009) 583e588. D. Zhang, X. Chen, J. Chi, J. Feng, Q. Wu, D. Zhu, ACS Catal. 5 (2015) 2452e2457. K.L. Kavanagh, M. Klimacek, B. Nidetzky, D.K. Wilson, Biochemistry 41 (2002) 8785e8795. R. Kratzer, B. Nidetzky, Chem. Commun. (2007) 1047e1049. R. Kratzer, S. Leitgeb, D.K. Wilson, B. Nidetzky, Biochem. J. 393 (2006) 51e58.
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi
I86A/C295A mutant secondary alcohol dehydrogenase from Thermoanaerobacter ethanolicus has broadened substrate specificity for aryl ketones Christopher M. Nealon a, 1, Travis P. Welsh a, 2, Chang Sup Kim b, Robert S. Phillips a, c, * a b c
Department of Chemistry, University of Georgia, Athens, GA 30602, USA Department of Chemical and Biological Engineering, Hanbat National University, Daejeon 34158, South Korea Department of Biochemistry and Molecular Biology, University of Georgia, Athens, GA 30602, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 29 June 2016 Received in revised form 29 July 2016 Accepted 1 August 2016 Available online 3 August 2016
Thermoanaerobacter ethanolicus secondary alcohol dehydrogenase (SADH) reduces aliphatic ketones according to Prelog's Rule, with binding pockets for small and large substituents. It was shown previously that the I86A mutant SADH reduces acetophenone, which is not a substrate of wild-type SADH, to give the anti-Prelog R-product (Musa, M. M.; Lott, N.; Laivenieks, M.; Watanabe, L.; Vieille, C.; Phillips, R. S. ChemCatChem 2009, 1, 89e93.). However, I86A SADH did not reduce aryl ketones with substituents larger than fluorine. We have now expanded the small pocket of the active site of I86A SADH by mutation of Cys-295 to alanine to allow reaction of substituted acetophenones. As predicted, the double mutant I86A/ C295A SADH has broadened substrate specificity for meta-substituted, but not para-substituted, acetophenones. However, the increase of the substrate specificity of I86A/C295A SADH is accompanied by a decrease in the kcat/Km values of acetophenones, possibly due to the substrates fitting loosely inside the more open active site. Nevertheless, I86A/C295A SADH gives high conversions and very high enantiomeric excess of the anti-Prelog R-alcohols from the tested substrates. © 2016 Elsevier Inc. All rights reserved.
Keywords: Mutagenesis Substrate specificity Stereospecificity Alcohol dehydrogenase Thermophilic
1. Introduction Two alcohol dehydrogenases, one with a specificity for primary alcohols, and a secondary alcohol dehydrogenase, were purified from Thermoanaerobacter ethanolicus, which was isolated from hot springs in Yellowstone National Park [1,2]. The wild-type secondary alcohol dehydrogenase (SADH) follows Prelog's rule in the reduction of ketones, adding the R-hydride of the NADPH to the re-face of the ketone [3]. The substrate specificity of wild-type SADH is restricted to primarily linear aliphatic and alicyclic ketones [4e9]. Fig. 1 shows the active site of SADH from Thermoanaerobacter brockii (formerly Thermoanaerobium brockii [10]), which is identical
Abbreviations: ADH, alcohol dehydrogenase; NADPþ, nicotinamide adenine dinucleotide phosphate; SADH, Thermoanaerobacter ethanolicus secondary ADH. * Corresponding author. Department of Chemistry, University of Georgia, Athens, GA 30602, USA E-mail address: [email protected] (R.S. Phillips). 1 Present address: Department of Chemistry, United States Naval Academy, Annapolis, MD 21402, USA. 2 Present address: Medical College of Georgia, Augusta University, Augusta, GA 30912, USA. http://dx.doi.org/10.1016/j.abb.2016.08.002 0003-9861/© 2016 Elsevier Inc. All rights reserved.
with the enzyme from T. ethanolicus [11]. Thus, a number of active site residues have been mutated in order to broaden the substrate specificity of SADH and make it useful in biocatalytic applications. These mutant SADHs, including S39T, C295A, W110A, and I86A, were found to have altered substrate specificity and stereoselectivity [9,12e14]. The I86A mutation of SADH not only expands substrate specificity to include acetophenone, which is a very poor substrate for wild-type SADH, but also reverses the usual preferred stereochemistry to produce the anti-Prelog R-product [14]. Unfortunately, I86A SADH exhibits limited reactivity with substituted acetophenones, fluorine being the only tolerable substituent found in the initial study. Homochiral 1-arylalkanols are useful intermediates in the preparation of pharmaceuticals. Thus, it was of interest to expand the active site further to allow the reduction of substituted acetophenones. The C295A mutant SADH was studied previously by Heiss and coworkers, and was found to increase the size of the alkyl group which can bind in the “small pocket” by one carbon atom [12]. Due to the proximity of Cys-295 to Ile-86 (Fig. 1), we predicted that having both mutations in the active site would expand the size of the small pocket to allow binding and reaction of
152
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156
Fig. 1. Stereoview of TbSADH, with residues of interest labeled. NADPþ shown in stick-form and zinc as a cyan sphere. This image was prepared with Pymol (The PyMOL Molecular €dinger, LLC) using the PDB file (1YKF). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version Graphics System, Version 1.7 Schro of this article.)
ring-substituted acetophenones with subsitutents larger than fluorine. To test this hypothesis, we have now prepared the I86A/ C295A double mutant of SADH, and we have found, as expected, that it has broadened specificity for substituted acetophenones. 2. Materials and methods
isolated from the host E. coli DH5a cells. Using the Quikchange method, the pADHB25-I86A plasmid was further mutated with the following forward and reverse primers, with the mutation indicated in bold. The resulting plasmid was sequenced to confirm the presence of the C295A mutation. C295A-F (ATAAAAGGCGGGCTAGCCCCCGGTGGACG). C295A-R (TTTCCGCCCGATCGGGGGCCACCTGCAGA).
2.1. General methods 2.4. Purification of secondary alcohol dehydrogenase Gas chromatography was performed with a Varian 3300 GC (Agilent Tech; Santa Clara, CA) using a Supelco (Sigma Aldrich; St. Louis, MO) b-Dex 120 cyclodextrin chiral column (30 m, 0.25 mm [i.d.], 0.25 mm film thickness) with He as the carrier gas and equipped with a flame ionization detector. Kinetic experiments and assays were performed on a Varian Cary 100 UVevisible spectrophotometer (Agilent Tech; Santa Clara, CA) equipped with a Peltier thermoelectric temperature-controlled 12-cell holder. 1H and 13C NMR analyses were collected on a Varian 400 MHz spectrometer with CDCl3 as the solvent at room temperature with tetramethylsilane or the solvent peak as the reference. 2.2. Materials Substrates were used as purchased from commercial suppliers with the exception of 3’-bromoacetophenone, which was prepared by a published procedure [15]. Acetophenone was purchased from Fisher Scientific (Waltham, MA). 20 ,40 -Difluoroacetophenone was bought from Acros (Geel, Belgium). The 2-acetylpyridine was purchased from Pfaltz and Bauer (waterbury, CT). The rest of the substrates tested were purchased from Sigma-Aldrich (St. Louis, MO). Acetonitrile was a product of Burdick and Jackson (Morristown, NJ). NADPH was purchased from Acros (Geel, Belgium). The antibiotics, kanamycin and ampicillin, were bought from Roche (Indianapolis, IN) and Fisher Biotech (fair lawn, NJ), respectively. Commercial grade solvents were used without further purification. The RedAagarose was obtained from Sigma-Aldrich (St. Louis, MO). 2.3. Mutagenesis The I86A plasmid, made as reported previously [14], was
The protein was expressed and purified based on our previously published method [9]. Modifications made to this method include the use of 50 mM Tris HCl, pH 8.0, buffer containing 5 mM DTT and 10 mM ZnCl2, a 5 mL RedA-agarose column, and optimization of the low salt (0.02 M NaClO4) and high salt (0.2 M NaClO4) wash and elution solutions. No further purification was needed after the RedA-agarose column as polyacrylamide gel electrophoresis showed a single band. The pooled fractions were stored at 80 C. 2.5. Enzyme assays The enzyme was assayed as previously described [14]. The enzyme assays were performed at 50 C in triplicate. The enzyme activity was measured in 50 mM potassium phosphate buffer (pH 6.5 at 50 C) with 0.4 mM NADPH to follow ketone reduction. The initial velocity was recorded on a Varian Cary 100 UV/Vis spectrophotometer at 340 nm for 10 min by monitoring NADPH consumption (ketone reduction). The enzyme was preincubated in 50 mM buffer solution with 5 mM DTT and 10 mM ZnCl2 at 50 C in the UV/Vis sample compartment for 10 min before addition to the assay mixtures. The substrate stock solutions were prepared in acetonitrile. Each assay contained not more than 5% acetonitrile, since higher concentrations were found to inhibit the enzyme activity. The kinetic data were fit to the Michaelis-Menten equation (Equation (1)) with the HYPERO program of Cleland [16]. The enzyme concentrations were determined from the A280 value of 0.82 for a 0.1% solution, and using the subunit molecular weight of 37.7 kDa.
v ¼ Vmax ½S=ðKm þ ½SÞ
(1)
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156
one carbon to propiophenone results in only a 2-fold decrease in kcat/Km. However, with increasing size of substituents on the ring, the kcat/Km value decreases dramatically. Interestingly, we found in this study that 3’-chloroacetophenone, which we had not tested previously, reacts with I86A mutant SADH, although the kcat/Km value is decreased by about 36-fold compared to acetophenone (Table 1). We had found previously that acetophenones with other substituents larger than fluorine are not detectable as substrates for I86A SADH, which was confirmed in this study (Table 1). Another interesting result is that 2-acetylpyridine, 4-acetylpyridine and 3acetylthiophene react much better than 3-acetylpyridine and 2acetylthiophene with I86A SADH. Due to the similar overall size of these heterocyclic substrates, these differences must be due to electronic effects rather than steric effects. The electronwithdrawing effects of the pyridine ring on the carbonyl are strongest in 2- and 4-acetylpyridine, due to resonance contributions, while the electron-donating effect of thiophene is weakest for 3-acetylthiophene. These observations suggest that hydride transfer to the carbonyl is at least partially rate-determining for these substrates and facilitated by making the carbonyl carbon more electron-deficient. The aromatic ring of these substrates must bind in the “small pocket” of I86A SADH, since the anti-Prelog R-enantiomer is the observed reduction product. We incorporated the C295A mutation into the I86A mutant SADH to expand the small pocket of the active site even further, and thus we hoped to broaden the scope of acetophenone substrates that can be reduced. While the I86A/C295A double mutant SADH unexpectedly showed no detectable reduction with para-substituted acetophenones, meta-substitution was accepted, as shown in Table 1. Small ortho-substituents are also tolerated, since 2’-methylacetophenone was found to be a slow substrate (Table 1). It is interesting that the kcat/Km value for unsubstituted acetophenone is reduced about 8-fold for I86A/ C295A SADH compared to I86A SADH. This suggests that the decrease in substrate specificity is gained at the expense of reaction efficiency. In contrast, 3’-chloroacetophenone is about 3-fold better as a substrate for I86A/C295A SADH than for I86A SADH. The Km values for all acetophenones with I86A/C295A SADH were
2.6. Asymmetric reduction with I86A and I86A/C295A Reactions included in 1 mL, substrate (0.21 mmol), 1 mg NADPH, 0.5 mL Tris-HCl buffer (8.0 pH with 10 mM ZnCl2), 0.2 mL 2propanol, 1 mg DTT and I86A or I86A/C295A SADH (0.654 mg). The reactions were incubated for 24 h at 50 C in a shaker. After 24 h, the reactions were quenched by addition of dichloromethane. This organic extract could be used to quantify the percent conversions by gas chromatography. The alcohol products were identified by comparison with the standards obtained from NaBH4 reduction of the ketones. Acetylation of the reaction products with acetic anhydride and pyridine was performed [17]. The enantiomeric excess of the products was found by GC analysis after acetylation. The ee values were calculated from the peak areas of the R and Sacetates, using Equation (2).
ðAreaðRÞ AreaðSÞÞ=ðAreaðRÞ þ AreaðSÞÞ* 100
153
(2)
2.7. Hammett-Taft calculations The curve fitting tool (cftool) in MATLAB (Version R2014b) was used to fit the three sets of data (Hammett, Taft, and either log(kcat/ (kcat)0) or log((kcat/KM)/(kcat/KM)0)) to the equation z ¼ a*x þ b*y. Equation (3) fitting parameters: 1.013 (SSE), 0.1435 (R2), 0.02113 (Adj. R2), 0.3805 (RMSE). Equation (4) fitting parameters: 0.2519 (SSE), 0.6082 (R2), 0.5522 (Adj. R2), 0.1897 (RMSE). Equation (3)a ¼ 1.531 (2.595, 5.656) b ¼ 1.312 (0.3748, 2.25). Equation (4)a ¼ 1.52 (0.5461, 3.586) b ¼ 1.105 (0.6524, 1.558). 3. Results and discussion In our previous report on the substrate specificity of I86A SADH, we did not evaluate kinetic parameters for all of the aromatic substrates [14]. As shown in Table 1, I86A SADH has the highest specificity constant, kcat/Km, with unsubstituted acetophenone, consistent with our previous results. Extension of the alkyl chain by
Table 1 I86A and I86A/C295A SADH kinetic parameters. O R1
R2 NADPH
Substrate
OH
SADH R1
R2
NADP+
I86A/C295A
I86A
R1
R2
kcat (s1)
kcat (s1) (M1s1)
kcat (s1)
kcat/Km (M1s1)
CH3 C6H5 30 -ClC6H5 30 -BrC6H5 30 -CH3C6H5 30 -IC6H5 20 -CH3C6H5 2-pyridyl 3-pyridyl 4-pyridyl 2-thienyl 3-thienyl C6H5 2,4-F2C6H3 30 -CH3OC6H5 30 -CF3C6H5
CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH2CH3 CH3 CH3 CH3
35 ± 6 15.8 ± 1.6 12.3 ± 2.4 6.5 ± 1.1 2.0 ± 0.33 0.61 ± 0.06 0.18 ± 0.04 2.31 ± 0.23 N/Aa 13.1 ± 1.9 1.5 ± 0.2 N/Aa 23.3 ± 5.5 2.9 ± 0.7 0.08 ± 0.02 0.04 ± 0.008
1080 ± 110 1900 ± 220 1300 ± 250 650 ± 110 180 ± 28 160 ± 30 18 ± 3 260 ± 30 630 ± 14 2400 ± 400 260 ± 60 350 ± 60 1500 ± 150 230 ± 60 210 ± 40 25 ± 10
1.01 ± 0.16 8.2 ± 0.6 0.66 ± 0.09 <0.01 <0.01 <0.01 <0.01 6.9 ± 1.6 0.47 ± 0.15 1.7 ± 0.1 1.7 ± 0.2 14.9 ± 4.2 6.5 ± 5.5 2.2 ± 0.7 <0.01 <0.01
1330 ± 340 14900 ± 3100 410 ± 90 <1 <1 <1 <1 1600 ± 260 200 ± 55 1050 ± 101 510 ± 85 4300 ± 730 8100 ± 1700 620 ± 140 <1 <1
a
Substrate inhibition precluded determination of kcat.
154
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156
generally around 8e10 mM, with some exceptions, so the kcat value is most affected by substitution, as can be seen in Table 1. The reactivity of the acetophenones also decreases with increasing size of the meta-substituent, and the meta-substituent can be as large as iodo, methoxy or trifluoromethyl. Since it appeared that both electronic and steric effects influence the reactivity of acetophenone substrates, a dual-parameter Hammett-Taft analysis [18] was performed with the kcat and kcat/ Km data for I86A/C295A SADH in Table 1. The plots of the results are shown in Fig. 2. The three-dimensional fit to the general equation z ¼ ax þ by for each was as follows:
Log Kcat ðKcat Þ0 ¼ 1:531s þ 1:312Es
(3)
Log ðKcat =Km Þ Kcat ðKcat Þ0 ¼ 1:531s þ 1:312Es
(4)
The sensitivity factors r and d (a and b of the general equation above) define how sensitive the kinetic parameters are to Hammett s and Taft ES, respectively. From each of the best fit planes, there is similar impact from the electronic effects as well as the steric effects on the reaction. The value of r is identical for both kcat and kcat/ Km, 1.53 and 1.52, respectively, in Equations (3) and (4), showing that both kinetic parameters are affected similarly by electronic effects. This suggests that the hydride transfer step for acetophenone reduction is partially rate-determining for both kcat and kcat/ Km. The positive value of r indicates that electron-withdrawal increases substrate reactivity, consistent with the results from
acetylpyridines
and
acetylthiophenes
discussed
above.
The
d values, 1.312 and 1.105, are slightly less than the r values and not significantly different. This suggests that the larger substituents may not allow the substrate to adopt the optimal geometry for hydride transfer on binding to the active site. Another comparison of the efficiency of the two mutant enzymes is through preparative reactions, as we had done previously. Table 2 compares the previously published I86A SADH data [14], with two additional substrates, and the new I86A/C295A SADH data. These reactions contained 0.21 mmol of substrate, corresponding to 25.2 mg in the case of acetophenone. In general, the preparative reaction results correlate well with the kinetic results in Table 1. As expected, 3’-chloroacetophenone reacts much better with I86A/C295A SADH than with I86A SADH, giving conversions of 23% and 0.6%, respectively. 2-Acetylpyridine and 4-acetylpyridine gave fairly high conversion with I86A/C295A SADH, although slightly lower conversion than the previously published results with I86A [14]. One interesting observation was that the percent conversion with 2-acetylthiophene and 3-acetylthiophene rose with 2-acetylthiophene with I86A/C295A SADH, while the percent conversion fell with 3-acetylthiophene with I86A/C295A SADH. 30 Aminoacetophenone, 3’-nitroacetophenone, 3’-ethoxyacetophenone, and 4’-methylacetophenone did not show any detectable reduction product with I86A/C295A SADH, and thus would be too large for the active site of I86A SADH. In all cases where reaction was observed, we found the product to be the antiPrelog R-alcohol with >99% ee.
Fig. 2. I86A/C295A SADH Hammett-Taft Plot. Top: Plot of substitutent effects on kcat. Bottom: Plot of substituent effects of kcat/Km.
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156 Table 2 I86A and I86A/C295 SADH GC assays. O
OH
SADH R2
R1
R2
R1
NADPH
NADP+
O
OH SADH
Substrate
I86A/C295A SADH
I86A SADH
R1
R2
Conv. (%)
ee (%)
R/S
Conv. (%)
ee (%)
R/S
C6H5 30 -ClC6H5 30 -BrC6H5 30 -IC6H5 30 -CH3C6H5 20 -CH3C6H5 30 -CH3OC6H5 30 -CF3C6H5 2-pyridyl 3-pyridyl 4-pyridyl 2-thienyl 3-thienyl C6H5 2,4-F2C6H3
CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH2CH3 CH3
68 23 8 4 23 <1 5 3 53 75 98 27 39 31 4
>99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99 >99
R R R R R R R R R R R R R R R
47a 0.6 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 >99a 46a >99a 2.52 76a 60a 33a
98a >99 e e e e e e >99a >99a >99a >99 >99a >99a >99a
Ra R e e e e e e Ra Ra Ra R Ra Ra Ra
a
From Musa et al. [14].
Reetz et al. have also performed mutagenesis studies with SADH from T. brockii in order to change the substrate specificity and stereoselectivity [19]. For the reduction of axially chiral cyclohexanones, which are reduced by wild-type SADH with good yield but modest enantioselectivity (66% ee), they found that the W110T mutant SADH produced the R-alcohols with high conversions and >92% ee. Conversely, the I86A mutant SADH discussed in this paper provided the S-alcohols in some cases, with stereoselectivities ranging from modest to excellent. Recently, they used triple code saturation mutagenesis to optimize the reduction of small saturated heterocyclic ketones which are difficult to reduce [20]. The residues in the active site targeted for mutagenesis included Ala-85, Ile-86, Trp-110, Leu-294, and Cys-295. After an initial round of saturation mutagenesis of each of these sites, the best mutations were grouped into two sites for a second round of mutagenesis. Site A (A85/I86/L294/C295) was mutated using a V-N-L triple code, and site B (A85/I86/W110/L294) was mutated with a V-Q-L triple code. A double mutant, I86N/C295N, was identified from the site A library that reduces the ketones to the R-alcohols in with very high ee. A triple mutant, I86V/W110L/L294Q, was found in the site B library that provide the S-alcohols in 95% ee. Mutagenesis has been used to modify substrate specificity of other dehydrogenases. The carbonyl reductase from Candida parapsilosis (CPCR2) contains a zinc ion, as does SADH, and also has
155
activity with medium chain alcohols. Jakoblinnert and coworkers expanded the active site of CPCR2, and thus affected the substrate specificity [21]. Five residues were selected for site-saturation mutagenesis (Leu-55, Pro-92, Gly-118, Leu-119, Leu-262). Of the active mutants at those locations, only L119 M gave activity that was greater than wild-type CPCR2. Even though methionine and leucine have side chains of similar size, the lack of branching in the methionine side chain provided a less constraining CPCR2 active site. Zhu et al. obtained a crystal structure of the carbonyl reductase, and using the structure as a guide, prepared three mutants, F285A, W286A, and F285A/W286A [22]. All of these mutant enzymes showed a reversal of stereochemistry in the reduction of aryl ketones, giving the anti-Prelog products. Zhu and coworkers also studied Sporobolomyces salmonicolor carbonyl reductase (SsCR), with the goal to improve the activity of wild-type SsCR towards para-substituted acetophenones [23]. They focused their site-saturation mutagenesis on the Gln-245 residue as a result of docking studies. The mutant enzymes with the greatest improvement were Q245H, Q245P, and Q245L, all three of which interestingly switched the stereochemistry of the alcohol product from R to S. They utilized site-saturation mutagenesis on the Met-242 residue to probe the SsCR active site further, and found improved properties for some of the substrates tested [23]. Also, M242Y, M242D, M242C and M242G mutations reversed the stereochemistry of the alcohol product. Since both Met-242 and Gln245 mutations alter the substrate specificity separately, they opted to design a double mutant at those sites. M242L/Q245P and M242L/ Q245T showed the largest boost in ee [23]. Broadening the substrate scope to include benzophenones, Li and coworkers found that mutagenesis at the Gln-245 residue improved substrate specificity as well as providing the S-isomer [24]. Zhang and coworkers recently reported a mutant SsCR study with b-aminoketones [25]. While there was a little activity with M242F/Q245T and b-aminoacetophenone, a docking study followed by sitesaturation mutagenesis yielded useful mutations at Pro-170 and Leu-174. L174Y and L174W were the single mutations with the greatest activity against b-aminoacetophenone. In addition, they designed double mutant enzymes of Pro-170 and Leu-174, with the most interesting mutant SsCRs being P170R/L174Y, P170R/L174W, and P170H/L174W. When each of these mutated enzymes were tested against b-amino-2-acetylthiophene, only L174Y and P170R/ L174Y gave high ee for each of the two substrates [25]. Kavanagh and coworkers studied the active site of Candida tenuis xylose reductase (CtXR) [26], which was later found to react with a-ketoesters [27]. Through modeling studies, the Trp-23 appeared to conflict with the a-carbonyl of the substrates. The tryptophan was replaced by phenylalanine and then tyrosine using site-directed mutagenesis. The activity was increased for each of these mutants when compared with wild-type CtXR against the various a-ketoesters [27]. Kratzer and researchers studied the kinetics of acetophenone and other assorted ketones reduction with W23F, W23Y and wild type CtXR [28]. While the mutations decreased the activity with most of the ketones, oxopantoyl lactone
Fig. 3. Cartoon showing the binding of meta-substituted acetophenone in the small pocket of I86A/C295A SADH.
156
C.M. Nealon et al. / Archives of Biochemistry and Biophysics 606 (2016) 151e156
and ethyl benzoylformate interestingly increased in activity of four to five-fold and two to three-fold, respectively. Due to the mixed kinetic results, the researchers rationalized that the Trp-23 has a role in selective binding and the specificity of the enzyme [28]. 4. Conclusions Taken together, our present results and the previously published results demonstrate that rational mutagenesis and directed evolution are practical approaches to broaden substrate specificity of alcohol dehydrogenases for reduction of a wide range of nonphysiological substrates. The wild-type SADH active site cannot accommodate a phenyl moiety in the small pocket, since wild-type SADH shows negligible activity with acetophenone in our assays [14]. Thus, the observation that only a single mutation (I86A) could open the small pocket enough for a phenyl ring to fit inside and give a substrate with a kcat/Km value of >104 M1 s1 (Table 1) is remarkable. The “small pocket” of wild-type SADH allows binding of substituents only as large as propyl or isopropyl [4,6,12]. The I86A mutation removes three carbons from the small pocket, and now the substrate of I86A SADH can have a six-carbon phenyl substituent. Hence, one can consider the active site of these enzymes essentially as a rigid scaffold, the shape of which can be tailored by mutagenesis to accommodate a wide range of substrates. We have now shown that incorporation of an additional mutation (C295A) of a neighboring residue in the I86A SADH active site has expanded the pocket to allow reaction of m-substituted aryl ketones (Fig. 3). The substituents that are tolerated are sterically similar to or smaller than the sulfur atom that was removed by mutagenesis. It is noteworthy that in increasing the size of the active site through multiple mutations, the efficiency of the enzyme decreased even as the substrate pool was able to increase. This could be because the substrate has more freedom of motion within the expanded active site. In conclusion, this I86A/C295A double mutation provides a starting point for further expansion of the SADH active site to allow reaction of both m- and p-substituted aromatic ketones.
Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.abb.2016.08.002. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
[22] [23] [24] [25]
Acknowledgements [26]
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
[27] [28]
J.G. Zeikus, P.W. Hegge, M.A. Anderson, Arch. Microbiol. 122 (1979) 41e48. J. Wiegel, L.G. Ljungdahl, Arch. Microbiol. 128 (1981) 343e348. V. Prelog, Pure Appl. Chem. 9 (1964) 119e130. E. Keinan, E.K. Hafeli, K.K. Seth, R. Lamed, J. Am. Chem. Soc. 108 (1986) 162e169. V.T. Pham, R.S. Phillips, L.G. Ljungdahl, J. Am. Chem. Soc. 111 (1989) 1935e1936. V.T. Pham, R.S. Phillips, J. Am. Chem. Soc. 112 (1990) 3629e3632. C.S. Zheng, V.T. Pham, R.S. Phillips, Bio. Med. Chem. Letts. 2 (1992) 619e622. C. Zheng, V.T. Pham, R.S. Phillips, Catal. Today 22 (1994) 607e620. A.E. Tripp, D.S. Burdette, J.G. Zeikus, R.S. Phillips, J. Am. Chem. Soc. 120 (1998) 5137e5141. Y.E. Lee, M.K. Jain, C. Lee, J.G. Zeikus, Int. J. Syst. Evol. Microbiol. 43 (1993) 41e51. C.M. Nealon, M.M. Musa, J.M. Patel, R.S. Phillips, ACS Catal. 5 (2015) 2100e2114. C. Heiss, M. Laivenieks, J.G. Zeikus, R.S. Phillips, Bioorg. Med. Chem. 9 (2001) 1659e1666. M.M. Musa, K.I. Ziegelmann-Fjeld, C. Vieille, J.G. Zeikus, R.S. Phillips, J. Org. Chem. 72 (2007) 30e34. M.M. Musa, N. Lott, M. Laivenieks, L. Watanabe, C. Vieille, R.S. Phillips, ChemCatChem 1 (2009) 89e93. D.E. Pearson, H.W. Pope, W.W. Hargrove, Org. Syn. 40 (1960) 7. A. Ghanem, V. Schuring, Tetrahedron Asymmetry 14 (2003) 57e62. W.W. Cleland, Methods Enzymol. 63 (1979) 103e138. R.W. Taft Jr., In “Steric Effects in Organic Chemistry”, in: M.S. Newman (Ed.), John Wiley & Sons, Inc., New York, N. Y.,, 1965, p. 565. R. Agudo, G.D. Roiban, M.T. Reetz, J. Am. Chem. Soc. 135 (2012) 1665e1668. Z. Sun, R. Lonsdale, A. Ilie, G. Li, J. Zhou, M.T. Reetz, ACS Catal. 6 (2016) 1598e1605. A. Jakoblinnert, J. Wachtmeister, L. Schukur, A.V. Shivange, M. Bocola, M.B. Ansorge-Schumacher, U. Schwaneberg, Protein Eng. Des. Sel. 26 (2013) 291e298. D. Zhu, Y. Yang, S. Majkowicz, T.H. Pan, K. Kantardjieff, L. Hua, Org. Lett. 10 (2008) 525e528. H. Li, Y. Yang, D. Zhu, L. Hua, K. Kantardjieff, J. Org. Chem. 75 (2010) 7559e7564. H. Li, D. Zhu, L. Hua, E.R. Biehl, Adv. Synth. Catal. 351 (2009) 583e588. D. Zhang, X. Chen, J. Chi, J. Feng, Q. Wu, D. Zhu, ACS Catal. 5 (2015) 2452e2457. K.L. Kavanagh, M. Klimacek, B. Nidetzky, D.K. Wilson, Biochemistry 41 (2002) 8785e8795. R. Kratzer, B. Nidetzky, Chem. Commun. (2007) 1047e1049. R. Kratzer, S. Leitgeb, D.K. Wilson, B. Nidetzky, Biochem. J. 393 (2006) 51e58.