- Email: [email protected]
Fundamental differences in bioaffinity of amino acid dehydrogenases for N6- and S6-linked immobilized cofactors using kinetic-based enzyme-capture strategies
ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 338 (2005) 102–112 www.elsevier.com/locate/yabio
Fundamental differences in bioaffinity of amino acid dehydrogenases for N6- and S6-linked immobilized cofactors using kinetic-based enzyme-capture strategies Jessica Forde, Laura Oakey, Linda Jennings, Patricia Mulcahy* Department of Applied Biology and Chemistry, Institute of Technology, Kilkenny Road, Carlow, Ireland Received 2 September 2004 Available online 8 December 2004
Abstract Five different immobilized NAD+ derivatives were employed to compare the behavior of four amino acid dehydrogenases chromatographed using kinetic-based enzyme capture strategies (KBECS): S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked immobilized NAD+. The amino acid dehydrogenases were NAD+-dependent phenylalanine (EC 1.4.1.20), alanine (EC 1.4.1.1), and leucine (EC 1.4.1.9) dehydrogenases from various microbial species and NAD(P)+-dependent glutamate dehydrogenase from bovine liver (GDH; EC 1.4.1.3). KBECS for bovine heart L -lactate dehydrogenase (EC 1.1.1.27) and yeast alcohol dehydrogenase (EC 1.1.1.1) were also applied to assist in a preliminary assessment of the immobilized cofactor derivatives. Results confirm that the majority of the enzymes studied retained affinity for NAD+ immobilized through an N6 linkage, as opposed to an N1 linkage, replacement of the nitrogen with sulfur to produce an S6 linkage, or attachment of the cofactor through the C8 position or the pyrophosphate group of the cofactor. The one exception to this was the dual-cofactor-specific GDH from bovine liver, which showed no affinity for N6-linked NAD+ but was biospecifically adsorbed to S6-linked NAD+ derivatives in the presence of its soluble KBEC ligand. The molecular basis for this is discussed together with the implications for future development and application of KBECS. 2004 Elsevier Inc. All rights reserved. Keywords: Kinetic-based enzyme capture strategies; Immobilized NAD+; Phenylalanine dehydrogenase; Alanine dehydrogenase; Leucine dehydrogenase; Glutamate dehydrogenase
The fabrication of protein-detecting arrays composed of various immobilized protein-capture agents is an active area of research [1–4]. The development of powerful protein-detecting arrays requires the immobilization of high-affinity (KD in the 109 to 1012 M range) and specific ligands [3]. Small ligands would have advantages over macromolecular protein or nucleic acid ligands as protein-capture agents because they can be readily obtained by chemical synthesis, but small molecules rarely bind to their protein with an affinity comparable to that
of an antibody. We have demonstrated that this drawback can be overcome for certain important enzyme families through the introduction of class-specific kinetic-based enzyme capture strategies (KBECS).1 While our studies to date have focused primarily on NAD(P)+-dependent dehydrogenases and chromatographic-based strategies [5,6], it should be possible to extrapolate the knowledge gained to the eventual development of cofactor-based protein capture arrays
1
*
Corresponding author. Fax: +353 59 91 70500. E-mail address: [email protected] (P. Mulcahy).
0003-2697/$ - see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2004.11.016
Abbreviations used: KBECS, kinetic-based enzyme-capture strategies; CBPCA, cofactor-based protein-capture arrays; DH, dehydrogenase.
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
(CBPCA). This could potentially include CBPCA for ATP-dependent protein kinases and NAD(P)(H)-dependent enzymes, the former constituting one of the largest target families in the human genome (approximately 600 members) with key functions in signal transduction. The key to the KBEC approach is the use of soluble analogues of the enzymes-specific substrate to specifically and reversibly push the rate of dissociation of the enzyme (from the immobilized cofactor) toward zero. Auxiliary kinetic-based tactics such as the stripping-ligand tactic can then be applied to minimize competition of the immobilized ligand for the target proteins with native factors in cell extracts [6]. The kinetic basis for KBECS has been detailed by Irwin and Tipton [7] for multisubstrate enzymes possessing either sequential compulsory-ordered or random kinetic mechanisms. Clearly the stability, specificity, and binding properties of the immobilized cofactor in the absence and presence of KBECS are prime considerations. While it was initially envisaged that a single immobilized NAD+, NADP+, or ATP derivative
103
would be developed for use with KBECS, recent studies suggest that different enzymes demonstrate inherently different affinities for differently immobilized cofactor derivatives. The primary objective of the present study was to extend the KBEC approach to other NAD+-dependent dehydrogenases and, in so doing, gain additional insights into the ideal properties of immobilized cofactor derivatives employed for such applications. Four amino acid dehydrogenases were used for these investigations: NAD+-dependent phenylalanine (PheDH; EC 1.4.1.20), alanine (AlaDH; EC 1.4.1.1), and leucine (LeuDH; EC 1.4.1.9) dehydrogenases from various microbial species and NAD(P)+-dependent glutamate dehydrogenase from bovine liver (GDH; EC 1.4.1.3). Bovine heart L -lactate dehydrogenase (L-LDH; EC 1.1.1.27) and yeast alcohol dehydrogenase (YADH; EC 1.1.1.1) were also included for comparative studies. A total of five different classes of immobilized NAD+ derivatives were investigated: S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked immobilized NAD+ (Fig. 1).
Fig. 1. Structures of the immobilized NAD+ derivatives.
104
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
Materials and methods All materials were obtained from Sigma Chemical (Poole, Dorset, England). Enzyme assays All dehydrogenase assays were monitored spectrophotometrically at 340 nm by measuring the increase or decrease in absorbance at 340 nm. GDH activity was assayed at 30 C in 50 mM potassium phosphate (pH 7.4) containing 50 mM ammonium sulphate, 160 lM NAD(P)H, and 5 mM a-ketoglutarate. PheDH activity was assayed at 25 C in 100 mM glycine (pH 10.4) containing 0.1 M KCl, 0.5 mM NAD+, and 10 mM L -phenylalanine. AlaDH was assayed at 37 C in 100 mM glycine (pH 8.6) containing 0.25 mM NAD+ and 2 mM L -alanine. LeuDH assays were performed at 55 C in 100 mM glycine (pH 8.6) containing 0.1 M NaCl, 2 mM L -leucine, and 0.5 mM NAD+. LDH assays were carried out at 30 C in 50 mM potassium phosphate (pH 7.4) containing 0.5 mM pyruvate and 0.2 mM NADH. YADH activity was assayed at 30 C in 30 mM sodium pyrophosphate (pH 8.8) containing 0.8 mM NAD+ and 0.12 M ethanol. Kinetic studies with GDH, LeuDH, AlaDH, and PheDH Kinetic studies were performed on commercial GDH, LeuDH, AlaDH, and PheDH preparations to identify potential competitive inhibitors that would act as KBEC ligands. Assays were carried out in triplicate. Lineweaver–Burk double-reciprocal plots were constructed and the sets of kinetic data fitted to straight lines by linear regression analysis using a Macintosh DeltaGraph applications program. The resulting fits and kinetic constants were concurrently evaluated statistically, R2 being better than 0.99 and usually 1.0 for all of the kinetic data presented. Synthesis and analysis of S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked NAD+ ˚ spacers) were synS6-linked NAD+ derivatives (9-A thesized as described [8]. Pyrophosphate-linked immobilized NAD+ derivative was prepared using carbodiimide-promoted direct coupling of NAD+ to a carboxylate-terminating spacer arm as described [9]. An 8 0 -azo-linked NAD+ derivative was synthesized with ˚ hydrophilic spacer arm using the method dea 14.8-A scribed [8]. N6-linked NAD+ derivatives were synthesized using a solid-phase modular approach involving the carbodiimide-promoted coupling of N6-carboxymethyl-NAD+ to two different spacer arm–Sepharose 4B ˚ ). assemblies (to produce spacer lengths of 9 and 19.5 A 6 + N -NAD derivatives were synthesized using 1,3-diamino-
˚ propanol as the spacer compound to produce the 9-A spacer length, while the extended polar spacer used to ˚ spacer length was prepared as deproduce the 19.5-A 6 scribed [10]. N -carboxymethyl-NAD+ was synthesized using the method of Lindberg et al. [11] but with some modifications recommended by Buchanan [12]: (i) alkylation of 1.0 g NAD+ with 3 g iodoacetic acid (rather than a 1:1 ratio); (ii) adjustment of the pH to pH 3.0 (rather than 3.5) following the 10-day incubation; (iii) Treatment of the crude N1-carboxymethyl-NAD+ with dithionite by dissolving the sample in 90 ml of 2% w/v sodium hydrogen carbonate (pH 8.5), deaerating with nitrogen for 2 min, adding 1.5 g sodium dithionite, and storing in the dark for 5 h before terminating the reaction by oxygenation for 12 min with air and for 3 min with nitrogen; (iv) rearrangement to N6-carboxymethyl-NAD+ by adjusting the pH to 11.5 and heating at 73–74 C for 2 h (rather than 1 h); and (v) enzymatic re-oxidation at room temperature for 6 h after the addition of 6 ml 2 M Tris, pH 7.5, 1.5 ml acetaldehyde, and YADH (11.7 mg, 5150 U). The crude N6-carboxymethyl-NAD+ was dissolved in 30 ml of distilled water (rather than 1 L) and the pH was adjusted to pH 8.0 with 1 M LiOH (rather than to pH 7.5 with NaOH). This solution was then applied to a Dowex 1 (·2), 200 mesh chloride form (dimensions, 5 · 12.5 cm), which had been previously washed with 4 column vol of 3 M HCl and equilibrated with distilled water. The solution was left to adsorb onto the column for 30 min before it was washed with 500 ml of distilled water and the starting buffer. A 2-L linear gradient was then applied (5 mM CaCl2, pH 2.7, to 100 mM CaCl2, pH 2.0; flow rate, 2.5 ml/min; fraction vol, 4 ml). N1carboxymethyl-NAD+ eluted at approximately 60 mM CaCl2 and N6-carboxymethyl-NAD+ eluted at approximately 80 mM CaCl2. All of the fractions collected were monitored at 260 nm and those containing the N6-NAD were pooled and lyophilized to give a white powder. This was desalted on a Sephadex G-10 column (2.5 · 19 cm) before being lyophilized once more to give a white product (stored at 70 C). The purified N6carboxymethyl-NAD+ was dissolved in distilled water to produce an absorbance reading of 90–450 at 266 nm. This was added to an equal vol of the spacer arm–Sepharose 4B complex. 1-Ethyl-3-(3-dimethylaminopropyl)-carbodiimide (148 mg/ml of matrix) was dissolved in 0.3 ml of water and 100 ll was added dropwise every 2 min. The pH was adjusted to pH 7.5 with 0.5 M HCl after each addition and for the following hour. The reaction was allowed to proceed for 30– 35 h at 19–21 C. The gel was then washed with 20 vol of 0.1 M NaCl and distilled water. N1-linked immobilized NAD+ derivatives were synthesized as generally described above, but the purified N6-carboxymethyl NAD+ was replaced with purified N1-carboxymethyl NAD+.
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
N1- and N6-linked immobilized NADP+ derivatives were used as control matrices to confirm the biospecificity of KBEC systems developed for the amino acid dehydrogenases with the N6-linked immobilized NAD+. These were prepared as described [5]. The total amount of immobilized NAD(P)+ present on the affinity matrices was ascertained using phosphorus analysis following chemical digestion of matrices [13]. Accessible immobilized NADP+ and NAD+ were determined by monitoring the enzymatic reduction of the immobilized cofactors by bovine liver GDH or by YADH and aldehyde dehydrogenase, respectively [8]. Please note that the amount of accessible immobilized 8 0 -azo-NAD+ cannot be determined in this way since this matrix absorbs at 340 nm. Analytical affinity chromatographic procedures Reversible biospecific adsorption of the test enzymes to the immobilized cofactor derivatives, in the presence of a suitable soluble KBEC ligand, was assessed using chromatography performed at room temperature in miniature columns (1.7 cm internal diameter; 1 ml bed vol). It was first confirmed that the test enzyme demonstrated no detectable affinity for the matrix in the absence of the KBEC ligand or, if the test enzyme was adsorbed under such conditions, that it could be fully recovered using soluble 5 0 -
105
AMP as a competitive counter-ligand. The test enzyme was then applied to the matrix in the presence of KBEC ligand, which was then maintained in the column buffer for 4 column vol, followed by the KBEC ligand plus 10 mM 5 0 -AMP for 3 column vol, followed by just the KBEC ligand for 2 column vol, and finally 10 column vol of column buffer containing no KBEC ligand or 5 0 -AMP. All fractions were analyzed for the test enzyme activity using the assays described earlier. The enzyme was deemed to have locked-onto the matrix if it autoeluted from the matrix following omission of the KBEC ligand from the column buffer. The presence of the KBEC ligand and/or 5 0 -AMP in enzyme samples and column fractions did not interfere with the level of enzyme activity detected provided that the volume of enzyme sample used in assays produced a 30-fold dilution (e.g., 100 ll of enzyme in a total assay volume of 3 ml). Percentage enzyme recoveries were calculated for all runs and these were always 93–100%. Control chromatographic experiments were also performed to confirm that any reported lockingon of the test enzymes was biospecific. In such cases the control matrix for S6-linked immobilized NAD+ derivatives was a structurally analogous S6-linked AMP. The control matrix for N6-linked immobilized NAD+ was a structurally analogous N6-linked NADP+. Representative chromatographic profiles are presented in Fig. 2.
Fig. 2. Representative chromatographic results for AlaDH demonstrating its preference for N6-linked immobilized NAD+ (accessible substitution, 0.78 lmol/g wet wt.) rather than S6-linked NAD+ (accessible substitution, 0.66 lmol/g wet wt.) for KBECS. Chromatography was performed on N6linked NAD+ derivative (A and B) or S6-linked immobilized NAD+ (C and D) in the absence (A and C) or presence of (B and D) 1 mM D -cysteine as soluble KBEC ligand plus 10 mM 5 0 -AMP as stripping ligand. The column buffer was 100 mM glycine buffer, pH 8.6, with the various additions indicated by the horizontal lines. The result represented in (B) indicates strong reversible biospecific adsorption in the presence of D -cysteine (++++), while that represented in (D) indicates no significant reversible biospecific adsorption in the presence of D -cysteine ().
106
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
When chromatography was performed at pH 8.6, the derivatives were stored between runs in potassium phosphate (pH 7.4) containing 0.02% sodium azide to avoid prolonged exposure of the derivatives to an alkaline environment and to inhibit bacterial growth.
Results and discussion Overview of the immobilized NAD+ derivatives Substitution properties of the S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked immobilized NAD+ derivatives utilized in the present study are summarized in Table 1. KBECS for bovine heart L-LDH and YADH were also applied to assist in a preliminary assessment of the immobilized cofactor derivatives. The results of these studies are also summarized in Table 1. Reversible biospecific adsorption (or locking-on) of both L-LDH and YADH was consistently achieved when using S6-, pyrophosphate-, or 8 0 -azo-linked immobilized NAD+ in conjunction with oxalate or hydroxylamine as the KBEC ligands (Table 1). But only L-LDH could be locked onto N1-linked immobilized NAD+. The N6linked NAD+ derivatives gave some variable results; e.g., YADH showed no affinity for the N6-NAD+ derivative synthesized to a substitution level of 1.66 lmol accessible NAD+/g wet wt., irrespective of whether 100 mM hydroxylamine was included in the column
buffer. But when the chromatographic run was repeated on an N6-NAD+ derivative with the lower accessible substitution level of 0.84 lmol accessible NAD+/g wet wt., the majority of the YADH locked-on in the presence of up to 100 mM hydroxylamine and eluted only upon omission of the KBEC ligand from the column buffer (Table 1). This suggested initially that the longer spacer arm of the latter derivative was required for the yeast enzyme, particularly when a second, more highly substituted N6-NAD+ (2.1 lmol accessible NAD+/g wet wt.) also failed to be effective in the locking-on mode with YADH (Table 1). That this conclusion was not the full story was indicated by successful locking-on to new Ôshort-armedÕ N6-NAD+ derivatives with accessible substitution levels ranging between 0.78 and 1.19 lmol accessible NAD+/g wet wt. These apparent anomalies may be partially attributed to subtle variations in different N6-carboxymethyl NAD+ preparations. Early results suggest that minor differences in the protocol used to produce the N6-carboxymethyl NAD+ can produce variations in the kinetic locking-on results with the yeast enzyme. For example, if the N-carboxymethyl nucleotide preparation is stored frozen prior to purification using ion-exchange chromatography, the derivatives produced are normally less effective with the YADH. In contrast to YADH, L-LDH was successfully locked-on to all N6-NAD+ derivatives with the exception of those with very low substitution levels (e.g., 0.2 lmol/g wet wt.). Neither of the test enzymes
Table 1 Overview of the properties of S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked NAD+ ˚) Position of attachment Spacer length (A Total substitution Accessible substitution S6-linkeda Pyrophosphateb N1-linked 8 0 -Azo-linkedc N6-linkedd
a
9 9 15 9 14 9 9 9 9 9 9 19.5 19.5
Test KBEC systemsg
(lmol/g wet wt.)
(lmol/g wet wt.)
L-LDH/oxalatee
YADH/hydroxylaminee
0.70 9.00 1.10 1.50 NDf 0.24 0.87 1.60 1.22 1.86 3.04 0.64 0.89
0.30 0.66 0.50 0.70 NDf 0.23 0.78 0.91 1.19 1.66 2.10 0.54 0.84
++++ ++++ ++++ ++++ ++++ — ++++ ++++ ++++ ++++ ++++ — ++++
++++ ++++ ++++ — ++++ — ++++ — ++++ — — — ++++
See [22] for further details. See [9] for further details. c See [8] for further details. d The different substitution levels resulted from coupling different concentrations of N6-carboxymethyl-NAD+: in descending order these were 11 AU at 266 nm (coupling time, 35 h), 90 AU (35 h), 90 AU (30 h), 450 AU (32 h), 240 AU (35 h), 236 AU (34 h), 600 AU (30 h), and 240 AU (32 h). e Column buffers were as follows: L-LDH, 50 mM phosphate (pH 7.4) containing up to 0.5 M KCl and oxalate (1–100 mM); YADH, 100 mM sodium pyrophosphate (pH 8.8) or 30 mM sodium pyrophosphate (pH 8.8) containing 0.1 M KCl and hydroxylamine (50–100 mM); 10 mM 5 0 -AMP was used as stripping ligand in all cases. f Earlier studies reported on the difficulty with determining the substitution level of 8 0 -azo-linked immobilized cofactor derivatives. The absorbance reading at 340 nm (kmax) was 9.2 AU/g wet wt. g Complete reversible biospecific adsorption in the presence of the KBEC ligand (++++), no adsorption (—). b
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
locked-on to analogous N6-linked NADP+ derivatives. These results confirm that the locking-on of the enzymes to the N6-linked immobilized NAD+ derivatives was biospecific. Taking into account (i) the substitution levels of the most effective N6-NAD+ derivatives, (ii) the locking-on ligand concentrations required for completed promotion of biospecific adsorption and (iii) the overall chromatographic behavior of the test enzymes employed in the present study, there was a clear indication that the L-LDH and YADH enzymes showed a difference in affinity for cofactors immobilized via thiol-linkage as opposed to N6-linkage, with much stronger locking-on systems being possible with S6-linked immobilized NAD+. Nevertheless, the N6-NAD+ derivatives were investigated in the locking-on mode with the four amino dehydrogenase activities targeted in the present study, with some unexpected results (see below). Potential KBEC ligands for the amino acid dehydrogenases Potential KBEC/locking-on ligands were identified for the target amino acid dehydrogenases using kinetic studies (Table 2). Effective KBEC ligands are generally competitive inhibitors with respect to the enzymes-specific substrate [6]. It has been shown that the dissociation constant (Ki value) for the KBEC ligand dictates the effective concentration of soluble KBEC ligand for bioaffinity chromatography; the lower the Ki value, the lower the concentration of ligand required in the column buffer for biospecific adsorption. However, another important factor that influences the Ki value of the enzyme for the specific substrate analogue is the substitution level of the immobilized cofactor derivative; the
Table 2 Representative kinetic constants for PheDH, AlaDH, LeuDH, and GDHa
PheDH AlaDH LeuDH
GDHb a
pH
Km (mM substrate)
Ki (mM inhibitor)
8.6 8.6 7.4 8.6 7.4 8.6 8.5 7.4
0.13 1.63 1.96 3.10 0.99 0.95 0.24 1.01
0.03 0.71 1.88 1.98 4.00 4.10 0.07 0.09
(L -phenylalanine) (L -alanine) (L -alanine) (L -leucine) (L -leucine) (L -leucine) (L -glutamate) (L -glutamate)
(D -phenylalanine) (D -cysteine) (D -cysteine) (D -norvaline) (D -norvaline) (D -valine) (glutarate) (glutarate)
PheDH assays were carried out at 25 C in 100 mM glycine containing 0.1 M KCl and 0.5 mM NAD+. AlaDH assays were carried out at 37 C with 0.25 M NAD+ in 100 mM glycine (pH 8.6) or 100 mM glycine (pH 7.4) containing 0.1 M KCl. LeuDH assays were carried out at 50 C with 1 mM NAD+ in 100 mM glycine containing 0.1 M NaCl (8.6) or 50 mM potassium phosphate (pH 7.4) containing 0.1 M NaCl. GDH assays were carried out at 30 C with 1 mM NAD+ in 50 mM potassium phosphate (pH 7.4) or 50 mM Hepes (pH 8.5). b Kinetic constants for GDH first reported in [22].
107
target enzyme elutes at progressively higher concentrations of soluble KBEC ligand as the substitution level of the affinity adsorbent is decreased [6]. Model studies with NAD+-dependent amino acid dehydrogenases PheDH has a sequential ordered mechanism proceeding through the formation of a ternary complex of the enzyme with NAD+ and L -phenylalanine and with ammonia, phenylpyruvate, and NADH released in that order [14]. Therefore this enzyme should be receptive to kinetic locking-on strategies based on immobilized cofactors [6]. Comparative bioaffinity studies with PheDH, D -phenylalanine, and the immobilized NAD+ derivatives are summarized in Table 3. In direct contrast to the success of KBECS with L-LDH (Table 1), PheDH showed no affinity for thiol-, pyrophosphate-, N1-, or 8 0 azo-linked immobilized NAD+ and eluted in the protein breakthrough peak irrespective of whether D -phenylalanine (10–100 mM) was included in the column buffer. In general, concentrations of 1–10 mM were effective for locking-on PheDH to N6- linked immobilized NAD+ ˚ spacers and substitution levels between 0.78 with 9-A and 2.10 lmol accessible NAD+/g wet wt. matrix. However, 0.1 mM D -phenylalanine was sufficient to promote strong adsorption of the enzyme to N6-linked immobi˚ spacers and lower accessible lized NAD+ with 19.5-A immobilized ligand concentrations of 0.54 lmol NAD+/g wet wt. These results would suggest that the locking-on of PheDH is achievable at lower KBEC ligand concentrations if the spacer arm length is increased. The reversible biospecific adsorption of PheDH was influenced by the salt concentration of the column buffer. Generally the enzyme exhibited weaker affinity for the matrix when KCl was omitted and thus the concentration of KBEC ligand required to effectively lockon the enzyme had to be increased (Table 3). The reason for the stronger adsorption of the enzyme to the immobilized derivatives in the presence of the salt is probably twofold. First, the Ki value of PheDH for D -phenylalanine may be lower in the presence of KCl. [This could not be confirmed experimentally due to the very low level of PhDH activity when assayed in the absence of KCl.] Second, it is possible to reinforce weak adsorption by introducing marginal non-biospecific hydrophobic interactions by the addition of salt to buffers [6]. Weakening of the locking-on effect was also observed when the pH was lowered from pH 8.6 to 7.4 (Table 3). Again the kinetic basis for this could not be established due to the very low activity of the enzyme at the lower pH value. AlaDH from Bacillus subtilis also has a sequential ordered kinetic mechanism where NAD+ binds before L -alanine, and ammonia, pyruvate, and NADH are
108
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
Table 3 Bioaffinity chromatographic behavior of L-PheDH from Sporosarcina species on immobilized NAD+ using KBECS Matrix ˚) S NAD (9 A ˚) S6-NAD+ (9 A ˚) S6-NAD+ (9 A ˚) Pyrophosphate (15 A ˚) N1-NAD+ (9 A ˚) 8 0 -Azo-(14 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A 6
+
[NAD+]
pH
[KCl]
Chromatographic behaviora ([D -phenylalanine] mM)
0.30 0.66 0.66 0.50 0.70 N.D. 2.10 2.10 1.66 1.66 1.19 1.19 0.78 0.23 1.19 1.19 0.84 0.54 0.54
8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 7.4 7.4 8.6 8.6 8.6
0.0 0.0 0.2 0.2 0.2 0.0 0.0 0.2 0.0 0.2 0.0 0.2 0.2 0.2 0.2 0.0 0.2 0.2 0.0
— — — — — — ++++ ++++ ++++ ++++ ++ ++++ ++++ ++++ ++ ++ Irreversible adsorption ++++ ++++
(10) (10) (10) (10) (100) (10) (10) (1) (10) (0.1) (10) (1) (1) (1) (10) (10) (0) (0.1) (2)
Chromatography was performed at room temperature in 50 mM potassium phosphate (pH 7.4) or 100 mM glycine (pH 8.6) and indicated additions. a (++++) Indicates strong adsorption in the presence of KBEC ligand; (++) indicates weaker adsorption; (—) indicates no adsorption in the presence of KBEC ligand.
released in that order [15]. The results of comparative bioaffinity studies (Table 4) using AlaDH and D -cysteine as the soluble KBEC ligand confirmed a pattern of behavior similar to that seen with PheDH (Table 3). Regardless of the substitution level or the concentration of KBEC ligand included in the column buffer (up to 100 mM D -cysteine), AlaDH eluted from the thiollinked immobilized cofactor derivatives in the protein
breakthrough peak. This enzyme also failed to lock-on to pyrophosphate-, N1-, and 8 0 -azo-linked immobilized NAD+ (Table 4). In contrast, N6-linked immobilized NAD+ again proved to be effective at pH 8.6 with D -cysteine concentrations as low as 0.1 mM with the most highly substituted matrix. When chromatography was repeated in the presence of 0.2 M KCl, the concentration of D -cysteine required to effectively lock-on the
Table 4 Bioaffinity chromatographic behavior of L-AlaDH from B. subtilis on immobilized NAD+ derivatives using KBECS Matrix ˚) S NAD (9 A ˚) S6-NAD+ (9 A ˚) Pyrophosphate (15 A ˚) N1-NAD+ (9 A ˚) 8 0 -Azo-(14 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A 6
+
[NAD+]
pH
[KCl]
Chromatographic behavioura ([D -cysteine] mM)
0.30 0.66 0.50 0.70 N.D. 2.10 2.10 1.66 1.66 1.19 1.19 1.19 0.91 0.78 2.10 2.10 0.54 0.54
8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 7.4 8.6 8.6 7.4 7.4 8.6 8.6
0.0 0.0 0.2 0.2 0.0 0.0 0.2 0.0 0.2 0.0 0.2 0.0 0.0 0.0 0.2 0.0 0.2 0.0
— — — — — ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ — ++++ — ++++
(100) (100) (100) (100) (100) (0.1) (100) (1) (100) (1) (100) (100) (1) (1) (100) (100) (100) (1)
Chromatography was performed at room temperature in 50 mM potassium phosphate (pH 7.4) or 100 mM glycine (pH 8.6) and indicated additions. a (++++) Indicates strong adsorption in the presence of KBEC ligand; (—) indicates no adsorption in the presence of KBEC ligand.
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
AlaDH to the immobilized derivative increased to 100 mM (Table 4). This effect was also seen with GDH when the enzyme exhibited only marginal affinity for the immobilized adsorbent in the presence of 5 mM glutarate and 0.5 M KCl but biospecifically locked-on to the matrix when no KCl was added to the column buffer [6]. This observation with GDH was explained by kinetic studies revealing competitive inhibition by glutarate in the absence of KCl and uncompetitive inhibition with the inclusion of 0.5 M KCl [6]. It is possible that the presence of salt also influences the type/degree of inhibition observed between the AlaDH and the D -cysteine. But this possibility could not be examined as the inclusion of 0.2 M KCl in the AlaDH enzyme assay interfered with the activity to such an extent that relevant kinetic data could not be obtained. Weakening of the locking-on effect was also observed when the pH was lowered from pH 8.6 to 7.4 (Table 4), although reversible biospecific adsorption could still be achieved at higher KBEC ligand concentrations (100 mM). While the increase in Ki observed at pH 7.4 may go some way toward explaining this observation (1.88 mM as opposed to 0.71 mM at pH 8.6; see Table 2), compound affinity effects may also be in play with the N6-NAD+ derivatives. Kinetic studies on Bacillus LeuDH have confirmed that the oxidative deamination proceeds through a sequential ordered mechanism via the formation of a ternary complex of the enzyme with NAD+ and the L amino acid [16]. Product inhibition patterns for the reductive amination also indicate a sequential ordered ternary–binary kinetic mechanism in which NADH, 2oxoisocaproate, and ammonia bind to the enzyme in this order, followed by the release of leucine and NAD+ from the enzyme [17,18]. The results of comparative bioaffinity studies with LeuDH are summarized in
109
Table 5 using D -norvaline and D -valine. In keeping with the chromatographic behavior of PheDH and AlaDH, LeuDH could be locked-on to N6-linked cofactor derivatives but only in the presence of highly substituted matrices (2.10 lmol/g wet wt.) and high concentrations of D -norvaline (100 mM). D -Valine promoted much weaker locking-on, presumably due to the difference in Ki values with LeuDH (Table 2). Similarly, locking-on was weakened or abolished at the lower pH of 7.4 (Table 5). This difference can again be explained by the higher Ki value of LeuDH with D -norvaline at pH 7.4 compared to that at pH 8.6. Model studies with NAD(P)+-dependent GDH While the exact nature of the sequential kinetic mechanism of bovine liver GDH continues to be the subject of some debate, it is generally believed to follow a random ordered rapid equilibrium mechanism for the oxidative deamination reaction [19,20] and a partially ordered mechanism in the reverse direction where ammonia can bind only after 2-ketoglutarate [21]. While this might seem initially to preclude KBECS for this enzyme based on the locking-on tactic, it has been argued that a form of locking-on could be envisaged in the case of a random sequential mechanism, if the equilibrium of the reaction under the conditions used is such that the binding of the second substrate favors the binding of the coenzyme [7]. The results of comparative lockingon bioaffinity studies for GDH are summarized in Table 6 using glutarate as the soluble KBEC ligand. These results confirm that, while the bovine liver GDH was completely locked-on to S6-NAD+ derivatives in the presence of relatively low concentrations of glutarate as locking-on ligand (1–10 mM), the enzyme failed to show any affinity for pyrophosphate-, N1-, or N6-linked
Table 5 Bioaffinity chromatographic behaviour of L-LeuDH from Bacillus species on immobilized NAD+ derivatives using KBECS Matrix ˚) S -NAD (9 A ˚) S6-NAD+ (9 A ˚) Pyrophosphate (15 A ˚) N1-NAD+ (9 A ˚) 8 0 -azo-(14 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (19.5 A 6
+
[NAD+]
pH
[KCl]
Chromatographic behavioura
0.30 0.66 0.50 0.70 N.D. 2.10 2.10 1.19 1.19 1.19 0.78 0.78 0.78 0.54
8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 7.4 8.6 8.6 7.4 8.6
0.2 0.2 0.2 0.2 0.0 0.2 0.2 0.2 0.2 0.0 0.2 0.2 0.2 0.2
— — — — — ++++ ++ ++ — — ++++ ++ ++++ —
100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM
D -norvaline D -norvaline D -norvaline D -norvaline D -norvaline D -norvaline D -valine D -norvaline D -valine D -norvaline D -norvaline D -valine D -norvaline D -norvaline
Chromatography was performed at room temperature in 50 mM potassium phosphate (pH 7.4) or 100 mM glycine (pH 8.6). a (++++) indicates strong adsorption in the presence of KBEC ligand; (++) indicates weaker adsorption; (—) indicates no adsorption in the presence of KBEC ligand.
110
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
Table 6 Bioaffinity chromatographic behavior of GDH from bovine liver on immobilized NAD+ derivatives using KBECS Matrix
[NAD+]
pH
Chromatographic Behaviourb ([glutarate] mM)
˚ )a S6-NAD+ (9 A ˚) S6-NAD+ (9 A ˚ )a S6-NAD+ (9 A ˚) Pyrophosphate (15 A ˚) N1-NAD+ (9 A ˚) 8 0 -Azo-(14 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A
0.30 0.66 0.30 0.50 0.70 N.D. 2.10 1.66 1.19 0.91 0.78 0.23 0.54 0.84 0.54
7.4 7.4 8.5 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4
++++ ++++ ++++ — — ++++ — — — — — — — — —
(10) (1) (50) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100)
Chromatography was performed at room temperature in 50 mM potassium phosphate (pH 7.4) or 50 mM Hepes (pH 8–9.5) and indicated additions. a S6-linked NAD+ results first reported in [22]. b (++++) Indicates strong adsorption in the presence of KBEC ligand; (—) indicates no adsorption in the presence of KBEC ligand.
immobilized NAD+. As already reported [22], the increase in Ki with increasing pH (Table 2) required an increase in the concentration of glutarate to ensure complete biospecific adsorption. Control chromatographic experiments with a structurally analogous S6-linked immobilized AMP confirmed the biospecific nature of the adsorption; GDH exhibited no affinity for the immobilized AMP in the presence of glutarate. In contrast, the bovine liver enzyme exhibited no affinity for N6-linked immobilized NAD+ synthesized to substitution levels of between 2.10 and 0.23 lmol accessible NAD+/g wet wt. matrix when chromatography was performed in the presence of glutarate (Table 6). The failure of bovine GDH to lock-on to the N6linked immobilized NAD+ derivatives was surprising in view of the fact that it was catalytically active with these immobilized cofactors (evidenced by the high percentage accessibilities determined enzymatically). Furthermore, the affinity precipitation of bovine GDH has been reported using the Bis-NAD+ bifunctional nucleotide ½N 2 ; N 02 adipodihydrazido bis ðN 6 Þ carboxymethyl NADþ in the presence of glutarate [23–25]. Steric hindrance cannot be invoked to explain the successful locking-on of GDH to Bis-NAD+ (where ˚ spacer joins the two amino groups of the adea 17-A nines) and S6-linked immobilized NAD+ (containing a ˚ spacer) but the failure to lock-on to N6-linked 9-A ˚ spacimmobilized NAD+ containing either 9- or 19.5-A ers. The more likely explanation resides in compound affinity effects where weak biospecific adsorption is reinforced by the introduction of marginal nonbiospecific hydrophobic interactions. The difference in locking-on behavior of GDH on the immobilized S6- and N6-linked immobilized NAD+ derivatives could therefore lie in the
hydrophobic nature of the sulfur group of the S6-linked cofactor derivative, which reinforces the biospecific interactions leading to ternary complex formation, with additional non-biospecific interaction effects in the active site of the GDH enzyme. A similar explanation has been put forward to interpret the continued adsorption of yeast hexokinase post-locking-on with citrate/Nacetyl-glucosamine to S6-linked immobilized ATP but not with N6-linked immobilized ATP derivatives [26]. Importantly, the nonbiospecific reinforcement interactions do not come into play until the GDH enzyme is biospecifically locked-on to the immobilized cofactor and they do not eclipse the biospecific interactions (evident from the autoelution of the enzyme from the matrix post-locking-on). Similar compound affinity effects could also be envisaged with Bis-NAD+ based on the nature of its spacer compound and through stabilization of the enzyme aggregates formed with Bis-NAD+ by intermolecular nonbiospecific interactions between individual enzyme molecules. The report that the dual-cofactor-specific enzyme could be successfully locked-on to N6-linked NADP+ derivatives [5] suggests that such compound affinity effects are not necessary to reinforce biospecific adsorption when approached using the alternate NADP+ as immobilized cofactor.
Conclusion A comparative study of the chromatographic behavior of four amino acid dehydrogenases on five very different immobilized NAD+ derivatives confirms that the majority of the enzymes studied retained affinity for NAD+ immobilized through an N6 linkage, as opposed
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
111
Table 7 Overview of the effect of the position and nature of attachment of immobilized NAD+ derivatives on the chromatographic behavior of NAD+dehydrogenases using KBECS Position of attachment
Bacillus LeuDH
Bacillus AlaDH
Sporosarcina PheDH
Bovine GDH
Bovine L-LDH
Yeast ADH
S6-linked 8 0 -Azo-linked Pyrophosphate N1-linked N6-linked
— — — — ++
— — — — ++
— — — — ++++
++++ ++++ — — —
++++ ++++ ++++ ++++ ++++
++++ ++++ ++++ — Variable
(++++) Indicates strong adsorption in the presence of KBEC ligand; (++) indicates weaker adsorption; (—) indicates no adsorption in the presence of KBEC ligand. See Tables 3–6 for further details.
to an N1 linkage, replacement of the nitrogen with sulfur to produce an S6 linkage, or attachment of the cofactor through the C8 position or the pyrophosphate group of the cofactor (Table 7). In direct contrast to the three bacterial NAD+-dependent amino acid dehydrogenases, the dual-cofactor-specific GDH from bovine liver showed no affinity for N6-linked NAD+, but lockedon strongly and biospecifically to lowly substituted S6linked NAD+ derivatives in the presence of millimolar concentrations of glutarate as KBEC ligand. The preference of certain NAD+-dependent dehydrogenases for N6-linked NAD+ compared to that for analogous S6-linked immobilized derivatives is probably due to an inability to accommodate the more bulky S6linked attachment within the active site. Confirmation of this explanation will require molecular modeling investigations once the full three-dimensional structures of the relevant enzymes are available. Other differences in affinity for immobilized cofactor derivatives can probably be explained with regard to compound affinity effects. For example, the stronger affinity of GDH (bovine liver), YADH, and L-LDH for S6-linked derivatives compared to that for analogous N6-linked cofactor could be due to a combination of biospecific and nonbiospecific interactions introduced by the hydrophobic sulfur group of the thiol derivative. The latter could reinforce the biospecific interactions leading to ternary complex formation and produce much stronger enzyme adsorption to the matrix. This compound affinity effect does not appear to dominate the biospecific interaction with bovine GDH, YADH, or L-LDH. From the results presented here and related studies, further development of S6-linked and N6-linked immobilized NAD+, NADP+, and ATP derivatives are required. In addition to desirable compound affinity effects, S6-linked NAD+ derivatives were found to be more stable than N6 linked derivatives. While most of the cofactors immobilized via N6 linkage showed signs of deterioration within a few months, the thiol derivatives were stable for a minimum of 2 years postsynthesis. Deterioration of N6-linked NAD+ is accelerated by prolonged exposure to high pH values (pH 8.6 and 10.2). In contrast the S6-linked NAD+ derivatives showed no loss in chromatographic effectiveness at alkaline pH values.
The main advantage of N6-linked cofactors is their wider applicability compared to those of any other derivatives produced to date. But their relative instability under alkaline pH is a drawback. One potential solution to the stability problems at alkaline pH involves approaching the kinetic locking-on strategy from the other side of the dehydrogenase-catalyzed reaction through using N6-linked NAD(P)H derivatives. In contrast to the N6-linked NAD+, N6-linked NADH stored at pH 10.2 for 5 months showed no signs of deterioration and no changes in chromatographic effectiveness (unpublished work). This raises the possibility of approaching the kinetic locking-on from the NAD(P)H side of the reaction when enzymes require N6-linked cofactors and when the target enzymes require alkaline environments for locking-on.
Acknowledgments This work was funded by the Technological Sector Research Programme (National Development Plan 2000–2006) and the Higher Education Authority (Ireland) under their Programme for Research in Third Level Institutions (PRTLI Cycle 1).
References [1] M.F. Lopez, M.G. Pluskal, Protein micro- and macroarrays: digitising the proteome, J. Chromatogr. B 787 (2003) 19–27. [2] H. Zhu, M. Snyder, Protein chip technology, Curr. Opin. Chem. Biol. 7 (2003) 55–63. [3] G. Elia, M. Silacci, S. Scheurer, J. Scheuermann, D. Neri, Affinity-capture reagents for protein arrays, Trends Biotechnol. 20 (Suppl 12) (2002) S19–S22. [4] T. Kodadek, Development of protein-detecting microarrays and related devices, Trends Biochem. Sci. 27 (6) (2002) 295–300. [5] M. McMahon, J. Tynan, P. Mulcahy, Kinetic locking-on and auxiliary tactics for bioaffinity purification of NADP+-dependent dehydrogenases using N6-linked immobilized NADP+-derivatives: studies with mammalian and microbial glutamate dehydrogenases, Biotechnol. Bioeng. 81 (3) (2003) 356–369. [6] P. Mulcahy, M. OÕFlaherty, Prospective: the kinetic locking-on strategy and auxiliary tactics for bioaffinity purification of NAD(P)+-dependent dehydrogenases, Anal. Biochem. 299 (2001) 1–18.
112
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
[7] J.A. Irwin, K.F. Tipton, Affinity precipitation: a novel approach to protein purification, Essays Biochem. 29 (1995) 137–155. [8] M. OÕFlaherty, M. McMahon, P. Mulcahy, A kinetic locking-on strategy for bioaffinity purification; Further studies with alcohol dehydrogenase, Protein Express. Pur. 15 (1999) 127–145. [9] L. Oakey, P. Mulcahy, Immobilized cofactor derivatives for kinetic-based enzyme capture strategies: direct coupling of NAD(P)+, Anal. Biochem. 335 (2004) 316–325. [10] P. OÕCarra, S. Barry, T. Griffin, Interfering and complicating adsorption effects in bioaffinity chromatography, in: W.B. Jacoby, M. Wilcheck (Eds.), Methods in Enzymology, vol. XXXIV, Academic Press, New York, 1974, pp. 108–126. [11] M. Lindberg, P.O. Larsson, K. Mosbach, A new immobilized NAD+ analogue: its application in affinity chromatography and as a functioning coenzyme, Eur. J. Biochem. 40 (1973) 187–193. [12] H. Buchanan, MSc. Thesis, National University of Ireland, Dublin, 1988. [13] K. Mosbach, AMP and NAD+ as general ligands, in: W.B. Jacoby, M. Wilcheck (Eds.), Methods in Enzymology, vol. XXXIV, Academic Press, New York, 1974, pp. 229–242. [14] H. Misona, J. Yonezawa, S. Nagata, S. Nagaski, Purification and characterization of a dimeric phenylalanine dehydrogenase from Rhodococcus maris K-18, J. Bacteriol. 171 (1989) 30–36. [15] W.W. Cleland, C.E. Grimshaw, Kinetic mechanism of Bacillus subtilis L -alanine dehydrogenase, Biochemistry 20 (1981) 5650– 5655. [16] W.W. Cleland, Steady state kinetics, in: P.D. Boyer (Ed.), third ed., The Enzymes, vol. II, Academic Press, New York, 1971, pp. 1–43.
[17] S. Nagata, H. Misona, Gene cloning, purification and characterization of thermostable and halophilic leucine dehydrogenase from a halophilic thermophile, Bacillus lichenformis TSN9, Appl. Microbiol. Biotechnol. 44 (1995) 432–438. [18] T. Sekimoto, S.J. Matsuyama, H.F. Fischer, Evidence for lysine 80 as general base catalyst of leucine dehydrogenase, J. Biol. Chem. 268 (1993) 27039–27045. [19] J.C. Rife, W.W. Cleland, Determination of the chemical mechanism of glutamate dehydrogenase from pH studies, Biochemistry 19 (1980) 2328–2333. [20] J.C. Rife, W.W. Cleland, Kinetic mechanism of glutamate dehydrogenase, Biochemistry 19 (1980) 2321–2328. [21] M.J. Adams, Oxidoreductases—pyridine nucleotide-dependent enzymes, in: M.I. Page, A. Williams (Eds.), Enzyme Mechanisms, The Royal Society of Chemistry, London, 1987, pp. 477–503. [22] M. OÕFlaherty, M. McMahon, J. Forde, P. Mulcahy, The kinetic locking-on strategy for bioaffinity purification; further studies with bovine liver glutamate dehydrogenase, Protein Express. Pur. 16 (1999) 276–297. [23] M. OÕDonnell, Enzyme kinetic affinity chromatography correlations for D -lactate and malate dehydrogenases. Ph.D., Thesis, National University of Ireland, Galway, 1994. [24] P.O. Larsson, K. Mosbach, Affinity precipitation of enzymes, FEBS Lett. 98 (1979) 333–338. [25] S. Flygare, T. Griffin, P.O. Larsson, K. Mosbach, Affinity precipitation of dehydrogenases, Anal. Biochem. 133 (1983) 409–416. [26] P. Mulcahy, M. OÕFlaherty, L. Jennings, T. Griffin, Application of kinetic-based biospecific affinity chromatographic systems to ATP-dependent enzymes, Anal. Biochem. 309 (2) (2002) 279–292.
Fundamental differences in bioaffinity of amino acid dehydrogenases for N6- and S6-linked immobilized cofactors using kinetic-based enzyme-capture strategies Jessica Forde, Laura Oakey, Linda Jennings, Patricia Mulcahy* Department of Applied Biology and Chemistry, Institute of Technology, Kilkenny Road, Carlow, Ireland Received 2 September 2004 Available online 8 December 2004
Abstract Five different immobilized NAD+ derivatives were employed to compare the behavior of four amino acid dehydrogenases chromatographed using kinetic-based enzyme capture strategies (KBECS): S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked immobilized NAD+. The amino acid dehydrogenases were NAD+-dependent phenylalanine (EC 1.4.1.20), alanine (EC 1.4.1.1), and leucine (EC 1.4.1.9) dehydrogenases from various microbial species and NAD(P)+-dependent glutamate dehydrogenase from bovine liver (GDH; EC 1.4.1.3). KBECS for bovine heart L -lactate dehydrogenase (EC 1.1.1.27) and yeast alcohol dehydrogenase (EC 1.1.1.1) were also applied to assist in a preliminary assessment of the immobilized cofactor derivatives. Results confirm that the majority of the enzymes studied retained affinity for NAD+ immobilized through an N6 linkage, as opposed to an N1 linkage, replacement of the nitrogen with sulfur to produce an S6 linkage, or attachment of the cofactor through the C8 position or the pyrophosphate group of the cofactor. The one exception to this was the dual-cofactor-specific GDH from bovine liver, which showed no affinity for N6-linked NAD+ but was biospecifically adsorbed to S6-linked NAD+ derivatives in the presence of its soluble KBEC ligand. The molecular basis for this is discussed together with the implications for future development and application of KBECS. 2004 Elsevier Inc. All rights reserved. Keywords: Kinetic-based enzyme capture strategies; Immobilized NAD+; Phenylalanine dehydrogenase; Alanine dehydrogenase; Leucine dehydrogenase; Glutamate dehydrogenase
The fabrication of protein-detecting arrays composed of various immobilized protein-capture agents is an active area of research [1–4]. The development of powerful protein-detecting arrays requires the immobilization of high-affinity (KD in the 109 to 1012 M range) and specific ligands [3]. Small ligands would have advantages over macromolecular protein or nucleic acid ligands as protein-capture agents because they can be readily obtained by chemical synthesis, but small molecules rarely bind to their protein with an affinity comparable to that
of an antibody. We have demonstrated that this drawback can be overcome for certain important enzyme families through the introduction of class-specific kinetic-based enzyme capture strategies (KBECS).1 While our studies to date have focused primarily on NAD(P)+-dependent dehydrogenases and chromatographic-based strategies [5,6], it should be possible to extrapolate the knowledge gained to the eventual development of cofactor-based protein capture arrays
1
*
Corresponding author. Fax: +353 59 91 70500. E-mail address: [email protected] (P. Mulcahy).
0003-2697/$ - see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2004.11.016
Abbreviations used: KBECS, kinetic-based enzyme-capture strategies; CBPCA, cofactor-based protein-capture arrays; DH, dehydrogenase.
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
(CBPCA). This could potentially include CBPCA for ATP-dependent protein kinases and NAD(P)(H)-dependent enzymes, the former constituting one of the largest target families in the human genome (approximately 600 members) with key functions in signal transduction. The key to the KBEC approach is the use of soluble analogues of the enzymes-specific substrate to specifically and reversibly push the rate of dissociation of the enzyme (from the immobilized cofactor) toward zero. Auxiliary kinetic-based tactics such as the stripping-ligand tactic can then be applied to minimize competition of the immobilized ligand for the target proteins with native factors in cell extracts [6]. The kinetic basis for KBECS has been detailed by Irwin and Tipton [7] for multisubstrate enzymes possessing either sequential compulsory-ordered or random kinetic mechanisms. Clearly the stability, specificity, and binding properties of the immobilized cofactor in the absence and presence of KBECS are prime considerations. While it was initially envisaged that a single immobilized NAD+, NADP+, or ATP derivative
103
would be developed for use with KBECS, recent studies suggest that different enzymes demonstrate inherently different affinities for differently immobilized cofactor derivatives. The primary objective of the present study was to extend the KBEC approach to other NAD+-dependent dehydrogenases and, in so doing, gain additional insights into the ideal properties of immobilized cofactor derivatives employed for such applications. Four amino acid dehydrogenases were used for these investigations: NAD+-dependent phenylalanine (PheDH; EC 1.4.1.20), alanine (AlaDH; EC 1.4.1.1), and leucine (LeuDH; EC 1.4.1.9) dehydrogenases from various microbial species and NAD(P)+-dependent glutamate dehydrogenase from bovine liver (GDH; EC 1.4.1.3). Bovine heart L -lactate dehydrogenase (L-LDH; EC 1.1.1.27) and yeast alcohol dehydrogenase (YADH; EC 1.1.1.1) were also included for comparative studies. A total of five different classes of immobilized NAD+ derivatives were investigated: S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked immobilized NAD+ (Fig. 1).
Fig. 1. Structures of the immobilized NAD+ derivatives.
104
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
Materials and methods All materials were obtained from Sigma Chemical (Poole, Dorset, England). Enzyme assays All dehydrogenase assays were monitored spectrophotometrically at 340 nm by measuring the increase or decrease in absorbance at 340 nm. GDH activity was assayed at 30 C in 50 mM potassium phosphate (pH 7.4) containing 50 mM ammonium sulphate, 160 lM NAD(P)H, and 5 mM a-ketoglutarate. PheDH activity was assayed at 25 C in 100 mM glycine (pH 10.4) containing 0.1 M KCl, 0.5 mM NAD+, and 10 mM L -phenylalanine. AlaDH was assayed at 37 C in 100 mM glycine (pH 8.6) containing 0.25 mM NAD+ and 2 mM L -alanine. LeuDH assays were performed at 55 C in 100 mM glycine (pH 8.6) containing 0.1 M NaCl, 2 mM L -leucine, and 0.5 mM NAD+. LDH assays were carried out at 30 C in 50 mM potassium phosphate (pH 7.4) containing 0.5 mM pyruvate and 0.2 mM NADH. YADH activity was assayed at 30 C in 30 mM sodium pyrophosphate (pH 8.8) containing 0.8 mM NAD+ and 0.12 M ethanol. Kinetic studies with GDH, LeuDH, AlaDH, and PheDH Kinetic studies were performed on commercial GDH, LeuDH, AlaDH, and PheDH preparations to identify potential competitive inhibitors that would act as KBEC ligands. Assays were carried out in triplicate. Lineweaver–Burk double-reciprocal plots were constructed and the sets of kinetic data fitted to straight lines by linear regression analysis using a Macintosh DeltaGraph applications program. The resulting fits and kinetic constants were concurrently evaluated statistically, R2 being better than 0.99 and usually 1.0 for all of the kinetic data presented. Synthesis and analysis of S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked NAD+ ˚ spacers) were synS6-linked NAD+ derivatives (9-A thesized as described [8]. Pyrophosphate-linked immobilized NAD+ derivative was prepared using carbodiimide-promoted direct coupling of NAD+ to a carboxylate-terminating spacer arm as described [9]. An 8 0 -azo-linked NAD+ derivative was synthesized with ˚ hydrophilic spacer arm using the method dea 14.8-A scribed [8]. N6-linked NAD+ derivatives were synthesized using a solid-phase modular approach involving the carbodiimide-promoted coupling of N6-carboxymethyl-NAD+ to two different spacer arm–Sepharose 4B ˚ ). assemblies (to produce spacer lengths of 9 and 19.5 A 6 + N -NAD derivatives were synthesized using 1,3-diamino-
˚ propanol as the spacer compound to produce the 9-A spacer length, while the extended polar spacer used to ˚ spacer length was prepared as deproduce the 19.5-A 6 scribed [10]. N -carboxymethyl-NAD+ was synthesized using the method of Lindberg et al. [11] but with some modifications recommended by Buchanan [12]: (i) alkylation of 1.0 g NAD+ with 3 g iodoacetic acid (rather than a 1:1 ratio); (ii) adjustment of the pH to pH 3.0 (rather than 3.5) following the 10-day incubation; (iii) Treatment of the crude N1-carboxymethyl-NAD+ with dithionite by dissolving the sample in 90 ml of 2% w/v sodium hydrogen carbonate (pH 8.5), deaerating with nitrogen for 2 min, adding 1.5 g sodium dithionite, and storing in the dark for 5 h before terminating the reaction by oxygenation for 12 min with air and for 3 min with nitrogen; (iv) rearrangement to N6-carboxymethyl-NAD+ by adjusting the pH to 11.5 and heating at 73–74 C for 2 h (rather than 1 h); and (v) enzymatic re-oxidation at room temperature for 6 h after the addition of 6 ml 2 M Tris, pH 7.5, 1.5 ml acetaldehyde, and YADH (11.7 mg, 5150 U). The crude N6-carboxymethyl-NAD+ was dissolved in 30 ml of distilled water (rather than 1 L) and the pH was adjusted to pH 8.0 with 1 M LiOH (rather than to pH 7.5 with NaOH). This solution was then applied to a Dowex 1 (·2), 200 mesh chloride form (dimensions, 5 · 12.5 cm), which had been previously washed with 4 column vol of 3 M HCl and equilibrated with distilled water. The solution was left to adsorb onto the column for 30 min before it was washed with 500 ml of distilled water and the starting buffer. A 2-L linear gradient was then applied (5 mM CaCl2, pH 2.7, to 100 mM CaCl2, pH 2.0; flow rate, 2.5 ml/min; fraction vol, 4 ml). N1carboxymethyl-NAD+ eluted at approximately 60 mM CaCl2 and N6-carboxymethyl-NAD+ eluted at approximately 80 mM CaCl2. All of the fractions collected were monitored at 260 nm and those containing the N6-NAD were pooled and lyophilized to give a white powder. This was desalted on a Sephadex G-10 column (2.5 · 19 cm) before being lyophilized once more to give a white product (stored at 70 C). The purified N6carboxymethyl-NAD+ was dissolved in distilled water to produce an absorbance reading of 90–450 at 266 nm. This was added to an equal vol of the spacer arm–Sepharose 4B complex. 1-Ethyl-3-(3-dimethylaminopropyl)-carbodiimide (148 mg/ml of matrix) was dissolved in 0.3 ml of water and 100 ll was added dropwise every 2 min. The pH was adjusted to pH 7.5 with 0.5 M HCl after each addition and for the following hour. The reaction was allowed to proceed for 30– 35 h at 19–21 C. The gel was then washed with 20 vol of 0.1 M NaCl and distilled water. N1-linked immobilized NAD+ derivatives were synthesized as generally described above, but the purified N6-carboxymethyl NAD+ was replaced with purified N1-carboxymethyl NAD+.
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
N1- and N6-linked immobilized NADP+ derivatives were used as control matrices to confirm the biospecificity of KBEC systems developed for the amino acid dehydrogenases with the N6-linked immobilized NAD+. These were prepared as described [5]. The total amount of immobilized NAD(P)+ present on the affinity matrices was ascertained using phosphorus analysis following chemical digestion of matrices [13]. Accessible immobilized NADP+ and NAD+ were determined by monitoring the enzymatic reduction of the immobilized cofactors by bovine liver GDH or by YADH and aldehyde dehydrogenase, respectively [8]. Please note that the amount of accessible immobilized 8 0 -azo-NAD+ cannot be determined in this way since this matrix absorbs at 340 nm. Analytical affinity chromatographic procedures Reversible biospecific adsorption of the test enzymes to the immobilized cofactor derivatives, in the presence of a suitable soluble KBEC ligand, was assessed using chromatography performed at room temperature in miniature columns (1.7 cm internal diameter; 1 ml bed vol). It was first confirmed that the test enzyme demonstrated no detectable affinity for the matrix in the absence of the KBEC ligand or, if the test enzyme was adsorbed under such conditions, that it could be fully recovered using soluble 5 0 -
105
AMP as a competitive counter-ligand. The test enzyme was then applied to the matrix in the presence of KBEC ligand, which was then maintained in the column buffer for 4 column vol, followed by the KBEC ligand plus 10 mM 5 0 -AMP for 3 column vol, followed by just the KBEC ligand for 2 column vol, and finally 10 column vol of column buffer containing no KBEC ligand or 5 0 -AMP. All fractions were analyzed for the test enzyme activity using the assays described earlier. The enzyme was deemed to have locked-onto the matrix if it autoeluted from the matrix following omission of the KBEC ligand from the column buffer. The presence of the KBEC ligand and/or 5 0 -AMP in enzyme samples and column fractions did not interfere with the level of enzyme activity detected provided that the volume of enzyme sample used in assays produced a 30-fold dilution (e.g., 100 ll of enzyme in a total assay volume of 3 ml). Percentage enzyme recoveries were calculated for all runs and these were always 93–100%. Control chromatographic experiments were also performed to confirm that any reported lockingon of the test enzymes was biospecific. In such cases the control matrix for S6-linked immobilized NAD+ derivatives was a structurally analogous S6-linked AMP. The control matrix for N6-linked immobilized NAD+ was a structurally analogous N6-linked NADP+. Representative chromatographic profiles are presented in Fig. 2.
Fig. 2. Representative chromatographic results for AlaDH demonstrating its preference for N6-linked immobilized NAD+ (accessible substitution, 0.78 lmol/g wet wt.) rather than S6-linked NAD+ (accessible substitution, 0.66 lmol/g wet wt.) for KBECS. Chromatography was performed on N6linked NAD+ derivative (A and B) or S6-linked immobilized NAD+ (C and D) in the absence (A and C) or presence of (B and D) 1 mM D -cysteine as soluble KBEC ligand plus 10 mM 5 0 -AMP as stripping ligand. The column buffer was 100 mM glycine buffer, pH 8.6, with the various additions indicated by the horizontal lines. The result represented in (B) indicates strong reversible biospecific adsorption in the presence of D -cysteine (++++), while that represented in (D) indicates no significant reversible biospecific adsorption in the presence of D -cysteine ().
106
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
When chromatography was performed at pH 8.6, the derivatives were stored between runs in potassium phosphate (pH 7.4) containing 0.02% sodium azide to avoid prolonged exposure of the derivatives to an alkaline environment and to inhibit bacterial growth.
Results and discussion Overview of the immobilized NAD+ derivatives Substitution properties of the S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked immobilized NAD+ derivatives utilized in the present study are summarized in Table 1. KBECS for bovine heart L-LDH and YADH were also applied to assist in a preliminary assessment of the immobilized cofactor derivatives. The results of these studies are also summarized in Table 1. Reversible biospecific adsorption (or locking-on) of both L-LDH and YADH was consistently achieved when using S6-, pyrophosphate-, or 8 0 -azo-linked immobilized NAD+ in conjunction with oxalate or hydroxylamine as the KBEC ligands (Table 1). But only L-LDH could be locked onto N1-linked immobilized NAD+. The N6linked NAD+ derivatives gave some variable results; e.g., YADH showed no affinity for the N6-NAD+ derivative synthesized to a substitution level of 1.66 lmol accessible NAD+/g wet wt., irrespective of whether 100 mM hydroxylamine was included in the column
buffer. But when the chromatographic run was repeated on an N6-NAD+ derivative with the lower accessible substitution level of 0.84 lmol accessible NAD+/g wet wt., the majority of the YADH locked-on in the presence of up to 100 mM hydroxylamine and eluted only upon omission of the KBEC ligand from the column buffer (Table 1). This suggested initially that the longer spacer arm of the latter derivative was required for the yeast enzyme, particularly when a second, more highly substituted N6-NAD+ (2.1 lmol accessible NAD+/g wet wt.) also failed to be effective in the locking-on mode with YADH (Table 1). That this conclusion was not the full story was indicated by successful locking-on to new Ôshort-armedÕ N6-NAD+ derivatives with accessible substitution levels ranging between 0.78 and 1.19 lmol accessible NAD+/g wet wt. These apparent anomalies may be partially attributed to subtle variations in different N6-carboxymethyl NAD+ preparations. Early results suggest that minor differences in the protocol used to produce the N6-carboxymethyl NAD+ can produce variations in the kinetic locking-on results with the yeast enzyme. For example, if the N-carboxymethyl nucleotide preparation is stored frozen prior to purification using ion-exchange chromatography, the derivatives produced are normally less effective with the YADH. In contrast to YADH, L-LDH was successfully locked-on to all N6-NAD+ derivatives with the exception of those with very low substitution levels (e.g., 0.2 lmol/g wet wt.). Neither of the test enzymes
Table 1 Overview of the properties of S6-, N6-, N1-, 8 0 -azo-, and pyrophosphate-linked NAD+ ˚) Position of attachment Spacer length (A Total substitution Accessible substitution S6-linkeda Pyrophosphateb N1-linked 8 0 -Azo-linkedc N6-linkedd
a
9 9 15 9 14 9 9 9 9 9 9 19.5 19.5
Test KBEC systemsg
(lmol/g wet wt.)
(lmol/g wet wt.)
L-LDH/oxalatee
YADH/hydroxylaminee
0.70 9.00 1.10 1.50 NDf 0.24 0.87 1.60 1.22 1.86 3.04 0.64 0.89
0.30 0.66 0.50 0.70 NDf 0.23 0.78 0.91 1.19 1.66 2.10 0.54 0.84
++++ ++++ ++++ ++++ ++++ — ++++ ++++ ++++ ++++ ++++ — ++++
++++ ++++ ++++ — ++++ — ++++ — ++++ — — — ++++
See [22] for further details. See [9] for further details. c See [8] for further details. d The different substitution levels resulted from coupling different concentrations of N6-carboxymethyl-NAD+: in descending order these were 11 AU at 266 nm (coupling time, 35 h), 90 AU (35 h), 90 AU (30 h), 450 AU (32 h), 240 AU (35 h), 236 AU (34 h), 600 AU (30 h), and 240 AU (32 h). e Column buffers were as follows: L-LDH, 50 mM phosphate (pH 7.4) containing up to 0.5 M KCl and oxalate (1–100 mM); YADH, 100 mM sodium pyrophosphate (pH 8.8) or 30 mM sodium pyrophosphate (pH 8.8) containing 0.1 M KCl and hydroxylamine (50–100 mM); 10 mM 5 0 -AMP was used as stripping ligand in all cases. f Earlier studies reported on the difficulty with determining the substitution level of 8 0 -azo-linked immobilized cofactor derivatives. The absorbance reading at 340 nm (kmax) was 9.2 AU/g wet wt. g Complete reversible biospecific adsorption in the presence of the KBEC ligand (++++), no adsorption (—). b
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
locked-on to analogous N6-linked NADP+ derivatives. These results confirm that the locking-on of the enzymes to the N6-linked immobilized NAD+ derivatives was biospecific. Taking into account (i) the substitution levels of the most effective N6-NAD+ derivatives, (ii) the locking-on ligand concentrations required for completed promotion of biospecific adsorption and (iii) the overall chromatographic behavior of the test enzymes employed in the present study, there was a clear indication that the L-LDH and YADH enzymes showed a difference in affinity for cofactors immobilized via thiol-linkage as opposed to N6-linkage, with much stronger locking-on systems being possible with S6-linked immobilized NAD+. Nevertheless, the N6-NAD+ derivatives were investigated in the locking-on mode with the four amino dehydrogenase activities targeted in the present study, with some unexpected results (see below). Potential KBEC ligands for the amino acid dehydrogenases Potential KBEC/locking-on ligands were identified for the target amino acid dehydrogenases using kinetic studies (Table 2). Effective KBEC ligands are generally competitive inhibitors with respect to the enzymes-specific substrate [6]. It has been shown that the dissociation constant (Ki value) for the KBEC ligand dictates the effective concentration of soluble KBEC ligand for bioaffinity chromatography; the lower the Ki value, the lower the concentration of ligand required in the column buffer for biospecific adsorption. However, another important factor that influences the Ki value of the enzyme for the specific substrate analogue is the substitution level of the immobilized cofactor derivative; the
Table 2 Representative kinetic constants for PheDH, AlaDH, LeuDH, and GDHa
PheDH AlaDH LeuDH
GDHb a
pH
Km (mM substrate)
Ki (mM inhibitor)
8.6 8.6 7.4 8.6 7.4 8.6 8.5 7.4
0.13 1.63 1.96 3.10 0.99 0.95 0.24 1.01
0.03 0.71 1.88 1.98 4.00 4.10 0.07 0.09
(L -phenylalanine) (L -alanine) (L -alanine) (L -leucine) (L -leucine) (L -leucine) (L -glutamate) (L -glutamate)
(D -phenylalanine) (D -cysteine) (D -cysteine) (D -norvaline) (D -norvaline) (D -valine) (glutarate) (glutarate)
PheDH assays were carried out at 25 C in 100 mM glycine containing 0.1 M KCl and 0.5 mM NAD+. AlaDH assays were carried out at 37 C with 0.25 M NAD+ in 100 mM glycine (pH 8.6) or 100 mM glycine (pH 7.4) containing 0.1 M KCl. LeuDH assays were carried out at 50 C with 1 mM NAD+ in 100 mM glycine containing 0.1 M NaCl (8.6) or 50 mM potassium phosphate (pH 7.4) containing 0.1 M NaCl. GDH assays were carried out at 30 C with 1 mM NAD+ in 50 mM potassium phosphate (pH 7.4) or 50 mM Hepes (pH 8.5). b Kinetic constants for GDH first reported in [22].
107
target enzyme elutes at progressively higher concentrations of soluble KBEC ligand as the substitution level of the affinity adsorbent is decreased [6]. Model studies with NAD+-dependent amino acid dehydrogenases PheDH has a sequential ordered mechanism proceeding through the formation of a ternary complex of the enzyme with NAD+ and L -phenylalanine and with ammonia, phenylpyruvate, and NADH released in that order [14]. Therefore this enzyme should be receptive to kinetic locking-on strategies based on immobilized cofactors [6]. Comparative bioaffinity studies with PheDH, D -phenylalanine, and the immobilized NAD+ derivatives are summarized in Table 3. In direct contrast to the success of KBECS with L-LDH (Table 1), PheDH showed no affinity for thiol-, pyrophosphate-, N1-, or 8 0 azo-linked immobilized NAD+ and eluted in the protein breakthrough peak irrespective of whether D -phenylalanine (10–100 mM) was included in the column buffer. In general, concentrations of 1–10 mM were effective for locking-on PheDH to N6- linked immobilized NAD+ ˚ spacers and substitution levels between 0.78 with 9-A and 2.10 lmol accessible NAD+/g wet wt. matrix. However, 0.1 mM D -phenylalanine was sufficient to promote strong adsorption of the enzyme to N6-linked immobi˚ spacers and lower accessible lized NAD+ with 19.5-A immobilized ligand concentrations of 0.54 lmol NAD+/g wet wt. These results would suggest that the locking-on of PheDH is achievable at lower KBEC ligand concentrations if the spacer arm length is increased. The reversible biospecific adsorption of PheDH was influenced by the salt concentration of the column buffer. Generally the enzyme exhibited weaker affinity for the matrix when KCl was omitted and thus the concentration of KBEC ligand required to effectively lockon the enzyme had to be increased (Table 3). The reason for the stronger adsorption of the enzyme to the immobilized derivatives in the presence of the salt is probably twofold. First, the Ki value of PheDH for D -phenylalanine may be lower in the presence of KCl. [This could not be confirmed experimentally due to the very low level of PhDH activity when assayed in the absence of KCl.] Second, it is possible to reinforce weak adsorption by introducing marginal non-biospecific hydrophobic interactions by the addition of salt to buffers [6]. Weakening of the locking-on effect was also observed when the pH was lowered from pH 8.6 to 7.4 (Table 3). Again the kinetic basis for this could not be established due to the very low activity of the enzyme at the lower pH value. AlaDH from Bacillus subtilis also has a sequential ordered kinetic mechanism where NAD+ binds before L -alanine, and ammonia, pyruvate, and NADH are
108
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
Table 3 Bioaffinity chromatographic behavior of L-PheDH from Sporosarcina species on immobilized NAD+ using KBECS Matrix ˚) S NAD (9 A ˚) S6-NAD+ (9 A ˚) S6-NAD+ (9 A ˚) Pyrophosphate (15 A ˚) N1-NAD+ (9 A ˚) 8 0 -Azo-(14 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A 6
+
[NAD+]
pH
[KCl]
Chromatographic behaviora ([D -phenylalanine] mM)
0.30 0.66 0.66 0.50 0.70 N.D. 2.10 2.10 1.66 1.66 1.19 1.19 0.78 0.23 1.19 1.19 0.84 0.54 0.54
8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 7.4 7.4 8.6 8.6 8.6
0.0 0.0 0.2 0.2 0.2 0.0 0.0 0.2 0.0 0.2 0.0 0.2 0.2 0.2 0.2 0.0 0.2 0.2 0.0
— — — — — — ++++ ++++ ++++ ++++ ++ ++++ ++++ ++++ ++ ++ Irreversible adsorption ++++ ++++
(10) (10) (10) (10) (100) (10) (10) (1) (10) (0.1) (10) (1) (1) (1) (10) (10) (0) (0.1) (2)
Chromatography was performed at room temperature in 50 mM potassium phosphate (pH 7.4) or 100 mM glycine (pH 8.6) and indicated additions. a (++++) Indicates strong adsorption in the presence of KBEC ligand; (++) indicates weaker adsorption; (—) indicates no adsorption in the presence of KBEC ligand.
released in that order [15]. The results of comparative bioaffinity studies (Table 4) using AlaDH and D -cysteine as the soluble KBEC ligand confirmed a pattern of behavior similar to that seen with PheDH (Table 3). Regardless of the substitution level or the concentration of KBEC ligand included in the column buffer (up to 100 mM D -cysteine), AlaDH eluted from the thiollinked immobilized cofactor derivatives in the protein
breakthrough peak. This enzyme also failed to lock-on to pyrophosphate-, N1-, and 8 0 -azo-linked immobilized NAD+ (Table 4). In contrast, N6-linked immobilized NAD+ again proved to be effective at pH 8.6 with D -cysteine concentrations as low as 0.1 mM with the most highly substituted matrix. When chromatography was repeated in the presence of 0.2 M KCl, the concentration of D -cysteine required to effectively lock-on the
Table 4 Bioaffinity chromatographic behavior of L-AlaDH from B. subtilis on immobilized NAD+ derivatives using KBECS Matrix ˚) S NAD (9 A ˚) S6-NAD+ (9 A ˚) Pyrophosphate (15 A ˚) N1-NAD+ (9 A ˚) 8 0 -Azo-(14 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A 6
+
[NAD+]
pH
[KCl]
Chromatographic behavioura ([D -cysteine] mM)
0.30 0.66 0.50 0.70 N.D. 2.10 2.10 1.66 1.66 1.19 1.19 1.19 0.91 0.78 2.10 2.10 0.54 0.54
8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 7.4 8.6 8.6 7.4 7.4 8.6 8.6
0.0 0.0 0.2 0.2 0.0 0.0 0.2 0.0 0.2 0.0 0.2 0.0 0.0 0.0 0.2 0.0 0.2 0.0
— — — — — ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ — ++++ — ++++
(100) (100) (100) (100) (100) (0.1) (100) (1) (100) (1) (100) (100) (1) (1) (100) (100) (100) (1)
Chromatography was performed at room temperature in 50 mM potassium phosphate (pH 7.4) or 100 mM glycine (pH 8.6) and indicated additions. a (++++) Indicates strong adsorption in the presence of KBEC ligand; (—) indicates no adsorption in the presence of KBEC ligand.
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
AlaDH to the immobilized derivative increased to 100 mM (Table 4). This effect was also seen with GDH when the enzyme exhibited only marginal affinity for the immobilized adsorbent in the presence of 5 mM glutarate and 0.5 M KCl but biospecifically locked-on to the matrix when no KCl was added to the column buffer [6]. This observation with GDH was explained by kinetic studies revealing competitive inhibition by glutarate in the absence of KCl and uncompetitive inhibition with the inclusion of 0.5 M KCl [6]. It is possible that the presence of salt also influences the type/degree of inhibition observed between the AlaDH and the D -cysteine. But this possibility could not be examined as the inclusion of 0.2 M KCl in the AlaDH enzyme assay interfered with the activity to such an extent that relevant kinetic data could not be obtained. Weakening of the locking-on effect was also observed when the pH was lowered from pH 8.6 to 7.4 (Table 4), although reversible biospecific adsorption could still be achieved at higher KBEC ligand concentrations (100 mM). While the increase in Ki observed at pH 7.4 may go some way toward explaining this observation (1.88 mM as opposed to 0.71 mM at pH 8.6; see Table 2), compound affinity effects may also be in play with the N6-NAD+ derivatives. Kinetic studies on Bacillus LeuDH have confirmed that the oxidative deamination proceeds through a sequential ordered mechanism via the formation of a ternary complex of the enzyme with NAD+ and the L amino acid [16]. Product inhibition patterns for the reductive amination also indicate a sequential ordered ternary–binary kinetic mechanism in which NADH, 2oxoisocaproate, and ammonia bind to the enzyme in this order, followed by the release of leucine and NAD+ from the enzyme [17,18]. The results of comparative bioaffinity studies with LeuDH are summarized in
109
Table 5 using D -norvaline and D -valine. In keeping with the chromatographic behavior of PheDH and AlaDH, LeuDH could be locked-on to N6-linked cofactor derivatives but only in the presence of highly substituted matrices (2.10 lmol/g wet wt.) and high concentrations of D -norvaline (100 mM). D -Valine promoted much weaker locking-on, presumably due to the difference in Ki values with LeuDH (Table 2). Similarly, locking-on was weakened or abolished at the lower pH of 7.4 (Table 5). This difference can again be explained by the higher Ki value of LeuDH with D -norvaline at pH 7.4 compared to that at pH 8.6. Model studies with NAD(P)+-dependent GDH While the exact nature of the sequential kinetic mechanism of bovine liver GDH continues to be the subject of some debate, it is generally believed to follow a random ordered rapid equilibrium mechanism for the oxidative deamination reaction [19,20] and a partially ordered mechanism in the reverse direction where ammonia can bind only after 2-ketoglutarate [21]. While this might seem initially to preclude KBECS for this enzyme based on the locking-on tactic, it has been argued that a form of locking-on could be envisaged in the case of a random sequential mechanism, if the equilibrium of the reaction under the conditions used is such that the binding of the second substrate favors the binding of the coenzyme [7]. The results of comparative lockingon bioaffinity studies for GDH are summarized in Table 6 using glutarate as the soluble KBEC ligand. These results confirm that, while the bovine liver GDH was completely locked-on to S6-NAD+ derivatives in the presence of relatively low concentrations of glutarate as locking-on ligand (1–10 mM), the enzyme failed to show any affinity for pyrophosphate-, N1-, or N6-linked
Table 5 Bioaffinity chromatographic behaviour of L-LeuDH from Bacillus species on immobilized NAD+ derivatives using KBECS Matrix ˚) S -NAD (9 A ˚) S6-NAD+ (9 A ˚) Pyrophosphate (15 A ˚) N1-NAD+ (9 A ˚) 8 0 -azo-(14 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (19.5 A 6
+
[NAD+]
pH
[KCl]
Chromatographic behavioura
0.30 0.66 0.50 0.70 N.D. 2.10 2.10 1.19 1.19 1.19 0.78 0.78 0.78 0.54
8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 8.6 7.4 8.6 8.6 7.4 8.6
0.2 0.2 0.2 0.2 0.0 0.2 0.2 0.2 0.2 0.0 0.2 0.2 0.2 0.2
— — — — — ++++ ++ ++ — — ++++ ++ ++++ —
100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM 100 mM
D -norvaline D -norvaline D -norvaline D -norvaline D -norvaline D -norvaline D -valine D -norvaline D -valine D -norvaline D -norvaline D -valine D -norvaline D -norvaline
Chromatography was performed at room temperature in 50 mM potassium phosphate (pH 7.4) or 100 mM glycine (pH 8.6). a (++++) indicates strong adsorption in the presence of KBEC ligand; (++) indicates weaker adsorption; (—) indicates no adsorption in the presence of KBEC ligand.
110
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
Table 6 Bioaffinity chromatographic behavior of GDH from bovine liver on immobilized NAD+ derivatives using KBECS Matrix
[NAD+]
pH
Chromatographic Behaviourb ([glutarate] mM)
˚ )a S6-NAD+ (9 A ˚) S6-NAD+ (9 A ˚ )a S6-NAD+ (9 A ˚) Pyrophosphate (15 A ˚) N1-NAD+ (9 A ˚) 8 0 -Azo-(14 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (9 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A ˚) N6-NAD+ (19.5 A
0.30 0.66 0.30 0.50 0.70 N.D. 2.10 1.66 1.19 0.91 0.78 0.23 0.54 0.84 0.54
7.4 7.4 8.5 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4
++++ ++++ ++++ — — ++++ — — — — — — — — —
(10) (1) (50) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100) (100)
Chromatography was performed at room temperature in 50 mM potassium phosphate (pH 7.4) or 50 mM Hepes (pH 8–9.5) and indicated additions. a S6-linked NAD+ results first reported in [22]. b (++++) Indicates strong adsorption in the presence of KBEC ligand; (—) indicates no adsorption in the presence of KBEC ligand.
immobilized NAD+. As already reported [22], the increase in Ki with increasing pH (Table 2) required an increase in the concentration of glutarate to ensure complete biospecific adsorption. Control chromatographic experiments with a structurally analogous S6-linked immobilized AMP confirmed the biospecific nature of the adsorption; GDH exhibited no affinity for the immobilized AMP in the presence of glutarate. In contrast, the bovine liver enzyme exhibited no affinity for N6-linked immobilized NAD+ synthesized to substitution levels of between 2.10 and 0.23 lmol accessible NAD+/g wet wt. matrix when chromatography was performed in the presence of glutarate (Table 6). The failure of bovine GDH to lock-on to the N6linked immobilized NAD+ derivatives was surprising in view of the fact that it was catalytically active with these immobilized cofactors (evidenced by the high percentage accessibilities determined enzymatically). Furthermore, the affinity precipitation of bovine GDH has been reported using the Bis-NAD+ bifunctional nucleotide ½N 2 ; N 02 adipodihydrazido bis ðN 6 Þ carboxymethyl NADþ in the presence of glutarate [23–25]. Steric hindrance cannot be invoked to explain the successful locking-on of GDH to Bis-NAD+ (where ˚ spacer joins the two amino groups of the adea 17-A nines) and S6-linked immobilized NAD+ (containing a ˚ spacer) but the failure to lock-on to N6-linked 9-A ˚ spacimmobilized NAD+ containing either 9- or 19.5-A ers. The more likely explanation resides in compound affinity effects where weak biospecific adsorption is reinforced by the introduction of marginal nonbiospecific hydrophobic interactions. The difference in locking-on behavior of GDH on the immobilized S6- and N6-linked immobilized NAD+ derivatives could therefore lie in the
hydrophobic nature of the sulfur group of the S6-linked cofactor derivative, which reinforces the biospecific interactions leading to ternary complex formation, with additional non-biospecific interaction effects in the active site of the GDH enzyme. A similar explanation has been put forward to interpret the continued adsorption of yeast hexokinase post-locking-on with citrate/Nacetyl-glucosamine to S6-linked immobilized ATP but not with N6-linked immobilized ATP derivatives [26]. Importantly, the nonbiospecific reinforcement interactions do not come into play until the GDH enzyme is biospecifically locked-on to the immobilized cofactor and they do not eclipse the biospecific interactions (evident from the autoelution of the enzyme from the matrix post-locking-on). Similar compound affinity effects could also be envisaged with Bis-NAD+ based on the nature of its spacer compound and through stabilization of the enzyme aggregates formed with Bis-NAD+ by intermolecular nonbiospecific interactions between individual enzyme molecules. The report that the dual-cofactor-specific enzyme could be successfully locked-on to N6-linked NADP+ derivatives [5] suggests that such compound affinity effects are not necessary to reinforce biospecific adsorption when approached using the alternate NADP+ as immobilized cofactor.
Conclusion A comparative study of the chromatographic behavior of four amino acid dehydrogenases on five very different immobilized NAD+ derivatives confirms that the majority of the enzymes studied retained affinity for NAD+ immobilized through an N6 linkage, as opposed
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
111
Table 7 Overview of the effect of the position and nature of attachment of immobilized NAD+ derivatives on the chromatographic behavior of NAD+dehydrogenases using KBECS Position of attachment
Bacillus LeuDH
Bacillus AlaDH
Sporosarcina PheDH
Bovine GDH
Bovine L-LDH
Yeast ADH
S6-linked 8 0 -Azo-linked Pyrophosphate N1-linked N6-linked
— — — — ++
— — — — ++
— — — — ++++
++++ ++++ — — —
++++ ++++ ++++ ++++ ++++
++++ ++++ ++++ — Variable
(++++) Indicates strong adsorption in the presence of KBEC ligand; (++) indicates weaker adsorption; (—) indicates no adsorption in the presence of KBEC ligand. See Tables 3–6 for further details.
to an N1 linkage, replacement of the nitrogen with sulfur to produce an S6 linkage, or attachment of the cofactor through the C8 position or the pyrophosphate group of the cofactor (Table 7). In direct contrast to the three bacterial NAD+-dependent amino acid dehydrogenases, the dual-cofactor-specific GDH from bovine liver showed no affinity for N6-linked NAD+, but lockedon strongly and biospecifically to lowly substituted S6linked NAD+ derivatives in the presence of millimolar concentrations of glutarate as KBEC ligand. The preference of certain NAD+-dependent dehydrogenases for N6-linked NAD+ compared to that for analogous S6-linked immobilized derivatives is probably due to an inability to accommodate the more bulky S6linked attachment within the active site. Confirmation of this explanation will require molecular modeling investigations once the full three-dimensional structures of the relevant enzymes are available. Other differences in affinity for immobilized cofactor derivatives can probably be explained with regard to compound affinity effects. For example, the stronger affinity of GDH (bovine liver), YADH, and L-LDH for S6-linked derivatives compared to that for analogous N6-linked cofactor could be due to a combination of biospecific and nonbiospecific interactions introduced by the hydrophobic sulfur group of the thiol derivative. The latter could reinforce the biospecific interactions leading to ternary complex formation and produce much stronger enzyme adsorption to the matrix. This compound affinity effect does not appear to dominate the biospecific interaction with bovine GDH, YADH, or L-LDH. From the results presented here and related studies, further development of S6-linked and N6-linked immobilized NAD+, NADP+, and ATP derivatives are required. In addition to desirable compound affinity effects, S6-linked NAD+ derivatives were found to be more stable than N6 linked derivatives. While most of the cofactors immobilized via N6 linkage showed signs of deterioration within a few months, the thiol derivatives were stable for a minimum of 2 years postsynthesis. Deterioration of N6-linked NAD+ is accelerated by prolonged exposure to high pH values (pH 8.6 and 10.2). In contrast the S6-linked NAD+ derivatives showed no loss in chromatographic effectiveness at alkaline pH values.
The main advantage of N6-linked cofactors is their wider applicability compared to those of any other derivatives produced to date. But their relative instability under alkaline pH is a drawback. One potential solution to the stability problems at alkaline pH involves approaching the kinetic locking-on strategy from the other side of the dehydrogenase-catalyzed reaction through using N6-linked NAD(P)H derivatives. In contrast to the N6-linked NAD+, N6-linked NADH stored at pH 10.2 for 5 months showed no signs of deterioration and no changes in chromatographic effectiveness (unpublished work). This raises the possibility of approaching the kinetic locking-on from the NAD(P)H side of the reaction when enzymes require N6-linked cofactors and when the target enzymes require alkaline environments for locking-on.
Acknowledgments This work was funded by the Technological Sector Research Programme (National Development Plan 2000–2006) and the Higher Education Authority (Ireland) under their Programme for Research in Third Level Institutions (PRTLI Cycle 1).
References [1] M.F. Lopez, M.G. Pluskal, Protein micro- and macroarrays: digitising the proteome, J. Chromatogr. B 787 (2003) 19–27. [2] H. Zhu, M. Snyder, Protein chip technology, Curr. Opin. Chem. Biol. 7 (2003) 55–63. [3] G. Elia, M. Silacci, S. Scheurer, J. Scheuermann, D. Neri, Affinity-capture reagents for protein arrays, Trends Biotechnol. 20 (Suppl 12) (2002) S19–S22. [4] T. Kodadek, Development of protein-detecting microarrays and related devices, Trends Biochem. Sci. 27 (6) (2002) 295–300. [5] M. McMahon, J. Tynan, P. Mulcahy, Kinetic locking-on and auxiliary tactics for bioaffinity purification of NADP+-dependent dehydrogenases using N6-linked immobilized NADP+-derivatives: studies with mammalian and microbial glutamate dehydrogenases, Biotechnol. Bioeng. 81 (3) (2003) 356–369. [6] P. Mulcahy, M. OÕFlaherty, Prospective: the kinetic locking-on strategy and auxiliary tactics for bioaffinity purification of NAD(P)+-dependent dehydrogenases, Anal. Biochem. 299 (2001) 1–18.
112
Amnio acid dehydrogenases, KBECS, and immobilized cofactors / J. Forde et al. / Anal. Biochem. 338 (2005) 102–112
[7] J.A. Irwin, K.F. Tipton, Affinity precipitation: a novel approach to protein purification, Essays Biochem. 29 (1995) 137–155. [8] M. OÕFlaherty, M. McMahon, P. Mulcahy, A kinetic locking-on strategy for bioaffinity purification; Further studies with alcohol dehydrogenase, Protein Express. Pur. 15 (1999) 127–145. [9] L. Oakey, P. Mulcahy, Immobilized cofactor derivatives for kinetic-based enzyme capture strategies: direct coupling of NAD(P)+, Anal. Biochem. 335 (2004) 316–325. [10] P. OÕCarra, S. Barry, T. Griffin, Interfering and complicating adsorption effects in bioaffinity chromatography, in: W.B. Jacoby, M. Wilcheck (Eds.), Methods in Enzymology, vol. XXXIV, Academic Press, New York, 1974, pp. 108–126. [11] M. Lindberg, P.O. Larsson, K. Mosbach, A new immobilized NAD+ analogue: its application in affinity chromatography and as a functioning coenzyme, Eur. J. Biochem. 40 (1973) 187–193. [12] H. Buchanan, MSc. Thesis, National University of Ireland, Dublin, 1988. [13] K. Mosbach, AMP and NAD+ as general ligands, in: W.B. Jacoby, M. Wilcheck (Eds.), Methods in Enzymology, vol. XXXIV, Academic Press, New York, 1974, pp. 229–242. [14] H. Misona, J. Yonezawa, S. Nagata, S. Nagaski, Purification and characterization of a dimeric phenylalanine dehydrogenase from Rhodococcus maris K-18, J. Bacteriol. 171 (1989) 30–36. [15] W.W. Cleland, C.E. Grimshaw, Kinetic mechanism of Bacillus subtilis L -alanine dehydrogenase, Biochemistry 20 (1981) 5650– 5655. [16] W.W. Cleland, Steady state kinetics, in: P.D. Boyer (Ed.), third ed., The Enzymes, vol. II, Academic Press, New York, 1971, pp. 1–43.
[17] S. Nagata, H. Misona, Gene cloning, purification and characterization of thermostable and halophilic leucine dehydrogenase from a halophilic thermophile, Bacillus lichenformis TSN9, Appl. Microbiol. Biotechnol. 44 (1995) 432–438. [18] T. Sekimoto, S.J. Matsuyama, H.F. Fischer, Evidence for lysine 80 as general base catalyst of leucine dehydrogenase, J. Biol. Chem. 268 (1993) 27039–27045. [19] J.C. Rife, W.W. Cleland, Determination of the chemical mechanism of glutamate dehydrogenase from pH studies, Biochemistry 19 (1980) 2328–2333. [20] J.C. Rife, W.W. Cleland, Kinetic mechanism of glutamate dehydrogenase, Biochemistry 19 (1980) 2321–2328. [21] M.J. Adams, Oxidoreductases—pyridine nucleotide-dependent enzymes, in: M.I. Page, A. Williams (Eds.), Enzyme Mechanisms, The Royal Society of Chemistry, London, 1987, pp. 477–503. [22] M. OÕFlaherty, M. McMahon, J. Forde, P. Mulcahy, The kinetic locking-on strategy for bioaffinity purification; further studies with bovine liver glutamate dehydrogenase, Protein Express. Pur. 16 (1999) 276–297. [23] M. OÕDonnell, Enzyme kinetic affinity chromatography correlations for D -lactate and malate dehydrogenases. Ph.D., Thesis, National University of Ireland, Galway, 1994. [24] P.O. Larsson, K. Mosbach, Affinity precipitation of enzymes, FEBS Lett. 98 (1979) 333–338. [25] S. Flygare, T. Griffin, P.O. Larsson, K. Mosbach, Affinity precipitation of dehydrogenases, Anal. Biochem. 133 (1983) 409–416. [26] P. Mulcahy, M. OÕFlaherty, L. Jennings, T. Griffin, Application of kinetic-based biospecific affinity chromatographic systems to ATP-dependent enzymes, Anal. Biochem. 309 (2) (2002) 279–292.