Editor: Roger Buxton
Characterization of Mycobacterium tuberculosis diaminopimelic acid epimerase: paired cysteine residues are crucial for racemization
Article first published online: 4 FEB 2008
© 2008 Federation of European Microbiological Societies
FEMS Microbiology Letters
Volume 280, Issue 1, pages 57–63, March 2008
How to Cite
Usha, V., Dover, L. G., Roper, D. L. and Besra, G. S. (2008), Characterization of Mycobacterium tuberculosis diaminopimelic acid epimerase: paired cysteine residues are crucial for racemization. FEMS Microbiology Letters, 280: 57–63. doi: 10.1111/j.1574-6968.2007.01049.x
- Issue published online: 4 FEB 2008
- Article first published online: 4 FEB 2008
- Received 3 September 2007; accepted 22 November 2007.First published online February 2008.
- diaminopimelic acid;
Recently, the overproduction of Mycobacterium tuberculosis diaminopimelic acid (DAP) epimerase MtDapF in Escherichia coli using a novel codon alteration cloning strategy and the characterization of the purified enzyme was reported. In the present study, the effect of sulphydryl alkylating agents on the in vitro activity of M. tuberculosis DapF was tested. The complete inhibition of the enzyme by 2-nitro-5-thiocyanatobenzoate, 5,5′-dithio-bis(2-nitrobenzoic acid) and 1,2-benzisothiazolidine-3-one at nanomolar concentrations suggested that these sulphydryl alkylating agents modify functionally significant cysteine residues at or near the active site of the epimerase. Consequently, the authors extended the characterization of MtDapF by studying the role of the two strictly conserved cysteine residues. The putative catalytic residues Cys87 and Cys226 of MtDapF were replaced individually with both serine and alanine. Residual epimerase activity was detected for both the serine replacement mutants C87S and C226S in vitro. Kinetic analyses revealed that, despite a decrease in the KM value of the C87S mutant for DAP that presumably indicates an increase in nonproductive substrate binding, the catalytic efficiency of both serine substitution mutants was severely compromised. When either C87 or C226 were substituted with alanine, epimerase activity was not detected emphasizing the importance of both of these cysteine residues in catalysis.
Despite the availability of appropriate chemotherapy for approaching 60 years, tuberculosis remains an enormous and universal healthcare concern (Dye, 2006; Kaufmann & Parida, 2007). The lack of a universally effective vaccine (Barreto et al., 2006) and the emergence of multidrug-resistant (MDR) and extensively drug-resistant (XDR) strains of the Mycobacterium tuberculosis complex, the aetiological agents for the disease, compound the problem. There is an urgent need for the development of new antituberculosis vaccines and drugs. Additionally, the definition of unexplored essential enzymes that may represent effective nodes of mycobacterial physiology for chemical intervention must be pursued.
The integrity of the complex mycobacterial envelope is essential with even minor changes compromising intracellular survival and virulence (Gao et al., 2003; Bhatt et al., 2007) and inhibitors of cell wall biosynthesis, such as isoniazid and ethambutol, have been clinically exploited as specific antimycobacterial drugs (Winder et al., 1970; Takayama & Kilburn, 1989). More generally, the inhibition of cross-linking of bacterial peptidoglycan by β-lactam antibiotics has proven extremely successful. The physical strength of peptidoglycan is related to its architecture; linear glycan chains are cross-linked by short stem peptides that vary in structure according to taxonomy (Schleifer & Kandler, 1972). Although mycobacterial peptidoglycan is unremarkable, penicillins are not clinically useful in treating tuberculosis. During exponential growth, mycobacteria cross-link stem peptides between the third [meso-diaminopimelic acid (DAP)] residue and the fourth (d-Ala) of another (Schleifer & Kandler, 1972; Wietzerbin et al., 1974). However, on entering stationary phase, M. tuberculosis incorporates meso-DAPmeso-DAP linkages between stem peptides, a penicillin resistant mode of ligation (Wietzerbin et al., 1974). As mycobacterial infection is complicated by the bacterium's ability to persist by entering a dormant phase, this altered mode of peptidoglycan cross-linking is likely relevant here (Goffin & Ghuysen, 2002). Importantly, however, meso-DAP is essential for both types of mycobacterial peptidoglycan cross-linking and thus plays a key role in mycobacterial cell wall biosynthesis.
DAP epimerase (DapF) catalyses the interconversion of ll-DAP and meso(dl)-DAP (Wiseman & Nichols, 1984) and has been identified and characterized in Escherichia coli (Wiseman & Nichols, 1984; Richaud et al., 1987; Higgins et al., 1989), Haemophilus influenzae (Cirilli et al., 1998; Koo & Blanchard, 1999; Lloyd et al., 2004; Pillai et al., 2006) and Corynebacterium glutamicum (Hartmann et al., 2003). Two active site cysteine residues of H. influenzae DAP epimerase, Cys73 and Cys217, are critical components in a two base mechanism; an active site thiolate deprotonates the α-carbon while the neighbouring thiol acts as an acid, protonating the resulting carbanion intermediate (Cirilli et al., 1998; Koo & Blanchard, 1999).
The authors recently cloned and overexpressed M. tuberculosis dapF using a novel codon alteration strategy, to enable its purification and characterization (Usha et al., 2006). It was reported here that the effect of alkylating agents on the wild-type MtDapF activity, the biochemical characterization of mutants of MtDapF generated by site-directed mutagenesis to examine the significance of cysteine residues to catalysis.
Materials and methods
The QuikChange II XL site-directed mutagenesis kit was procured from Stratagene. Oligonucleotide primers were synthesized by MWG Biotech, Germany. A newly-synthesized batch of [2,6-3H] DAP, 60 Ci mmol−1 was obtained from American Radiolabeled Chemicals. Cation exchange resin AG 50W-X8, 200–400 mesh, H+ form was obtained from BioRad Laboratories. All other chemicals were of reagent grade and purchased from Sigma. Escherichia coli C41(DE3) (Imaxio, France) was used as a host for protein production. HisTrap Ni2+ Sepharose high performance 1-mL column was obtained from GE Healthcare.
DAP epimerase assay
The assay measures the release of 3H to H2O from [2,6-3H] DAP (Wiseman & Nichols, 1984; Richaud et al., 1987). DAP epimerase activity was assayed as described previously (Usha et al., 2006) with minor modifications. The reaction, in a final volume of 100 μL, contained 0.1 M Tris-HCl pH 7.5, 1 mM EDTA, 1 mM dithiothreitol, 0.5 μCi [2,6-3H] DAP, 64 μM dl-DAP and 125 ng of recombinant wild type M. tuberculosis DapF. The reactions were incubated at 30 °C for 30 min and terminated by addition of 10% tricarboxylic acid. The quenched reaction mixture was applied to a 2-mL AG 50W-X8 ion exchange column (H+ form) packed in a syringe, which was washed with 3 × 2 mL of water, the eluates were combined and radioactivity was quantified by liquid scintillation counting. The quantity of ll-DAP formed in each sample was determined after subtraction of 3H release from identical reactions without added enzyme from those containing the DapF. Where alkylating agents were used these were incubated with the enzyme at 37 °C for 10 min before initiating the assay through the addition of meso-DAP.
The variable concentrations of the wild-type enzyme used for the assay ranged from 32 to 250 ng and that for the C87S or C226S mutants ranged from 0.5 to 20 μg and C87A or C226A mutants ranged from 4 to 24 μg. All measurements were carried out in duplicate.
Mycobacterium tuberculosis DapF mutants were constructed by the individual replacement of two cysteine residues (cysteine 87 and 226) with either alanine or serine residues using QuikChange II XL site-directed mutagenesis kit (Stratagene) according to the manufacturers' instructions. The primers used for mutagenesis are described in Table 1. The pET 28b-dapFca vector (Usha et al., 2006) was used as the template for the mutagenic PCR. After PCR, the template plasmid DNA was digested with Dpn I and the nonmethylated mutated plasmid DNA was used to transform E. coli XL10-Gold. Transformants were cultured on Luria–Bertani (LB) agar containing 25 μg mL−1 kanamycin sulphate. Single transformant clones were cultured and plasmid DNA extracted. The validity of the constructs was confirmed via nucleotide sequencing.
|Mutation||Primer sequence (5′3′)|
Expression, purification and characterization of MtDapF mutants
The mutants of M. tuberculosis dapF were identically expressed like the wild-type gene in E. coli C41 (DE3) cells. A single colony was used to inoculate 10 mL overnight culture that subsequently provided the inoculum for 1 L cultures in Terrific broth containing 25 μg mL−1 kanamycin sulphate. The cultures were incubated at 37 °C until it reached an OD600 nm of 0.6, induced with 1 mM isopropyl-β-d-thiogalactopyranoside and incubated at 16 °C overnight. The cells were harvested and stored in −80 °C. The wild-type M. tuberculosis DapF and the mutant enzymes were purified as described previously (Usha et al., 2006) with minor modifications. Cell pellets were resuspended in binding buffer (20 mM HEPES pH 8.0, 500 mM NaCl, 50 mM imidazole, 10% glycerol, 5 mM dithiothreitol, 1 mM phenylmethanesulphonyl fluoride, 1 mM benzamidine and 1 mg mL−1 lysozyme) and incubated for 30 min at 4 °C. After lysis by French press, 5 μg mL−1 of DNAse and RNAse were added to the crude extract kept on ice for 10 min and centrifuged at 27 000 g for 45 min at 4 °C to remove cell debris and the supernatant applied to a 1 mL HisTrap Ni2+ Sepharose high performance column (GE Healthcare) equilibrated with binding buffer without the protease inhibitors and lysozyme. The column was extensively washed with binding buffer to remove contaminating proteins. MtDapF was recovered by elution with 2 mL of elution buffer containing 20 mM HEPES pH 8.0, 500 mM NaCl, 10% glycerol and 150 mM imidazole. The purified protein was dialysed against 20 mM HEPES pH 8.0, 10% glycerol, 10 mM dithiothreitol and 1 mM EDTA, concentrated using Centricon YM-10 filter units and stored as aliquots in −80 °C.
Determination of KM app for MtDapF mutants and wild-type enzyme
The kinetic parameter, apparent KM (KM app) of the mutant and wild-type enzyme was determined by measuring initial velocity in the presence of varying concentrations of meso-DAP. All values were obtained from duplicate readings. Kinetic constants, KM app and Vmax were obtained from Lineweaver–Burke plots.
In silico analyses of MtDapF primary structure
Alignment of the amino acyl residue sequence of MtDapF with the characterized DAP epimerases of H. influenzae, E. coli and C. glutamicum using the clustalw algorithm revealed a significant degree of sequence similarity; the mycobacterial enzyme shares 28% amino acid identity with its H. influenzae orthologue (Fig. 1a). The primary structure of MtDapF contains five cysteine residues, two of which are strictly conserved in all of its characterized orthologues (Fig. 1a) and are crucial for DAP epimerase activity (Koo & Blanchard, 1999; Koo et al., 2000). These two cysteine residues have been shown to be close neighbours in structural studies and readily form a disulphide linkage (Born & Blanchard, 1999). The authors' homology modelling studies (Fig. 1b) suggest that MtDapF likely adopts a very similar fold to that of the H. influenzae enzyme. In this model Cys residues 87 and 226 superimpose with their H. influenzae counterparts, which were determined as cystine (Lloyd et al., 2004). The potential for the formation of a similar disulphide linkage in MtDapF appears consistent with its often rapid oxidative inactivation (data not shown).
Inhibition of M. tuberculosis DAP epimerase with sulphydryl alkylating agents
The apparent conservation in MtDapF of catalytically active cysteine residues (Fig. 1) associated with pyridoxyl-5-phosphate (PLP)-independent amino acid racemization (Rudnick & Abeles, 1975; Gallo & Knowles, 1993; Tanner et al., 1993; Koo et al., 2000) led to the investigation of the effects upon catalysis of three commercially available sulphydryl alkylating agents, 2-nitro-5-thiocyanatobenzoate (NTCB), 5,5′-dithio-bis(2-nitrobenzoic acid) (DTNB) and 1,2-benzisothiazolidine 3-one (BIT). The Bacillus subtilis glutamate racemase was inactivated on incubation with 0.2 mM NTCB at 37 °C for 10 min (Ashiuchi et al., 1998) and DTNB rapidly and completely inactivates the Lactobacillus fermenti glutamate racemase (Gallo & Knowles, 1993). Here the panel of agents were tested as inhibitors of DAP epimerase activity by preincubating for 10 min at 37 °C with MtDapF over a range of concentrations before initiation of each assay by the addition of meso-DAP. Pretreatment with all three alkylating agents at 30 nM for ten minutes resulted in complete inhibition of the DAP epimerase activity of MtDapF (Fig. 2). The sulphydryl specificity of these alkylating agents strongly suggests an essential role for at least one cysteine residue in the catalytic mechanism.
DAP epimerase activity of M. tuberculosis DapF mutants
As part of this ongoing investigation of MtDapF as a potential novel antimycobacterial drug target and in order to assess whether these residues are important elements of the catalytic mechanism here, a series of four mutants of MtDapF by the individual replacement of these two cysteine residues (Cys87 and Cys226) with either alanine or serine residues were generated.
Like the wild-type allele, optimal expression of M. tuberculosis dapF mutants was achieved with 1 mM isopropyl β-d-1-thiogalactopyranoside at 16 °C for 24 h. The mutant proteins were purified in a single step by metal chelate affinity chromatography on Ni2+ Sepharose in an identical fashion to the wild-type enzyme. Sodium dodecyl sulphate-polyacrylamide gel electrophoresis of mutant enzymes showed a major band with mobility identical to that of the wild-type enzyme although relative yields of mutant protein recovered following purification were decreased (Fig. 3).
The DAP epimerase activity of purified wild-type M. tuberculosis DapF and mutants was assayed in vitro by following the release of 3H from 3H-DAP to 3H2O (Usha et al., 2006). Since a freshly synthesized batch of 3H-DAP was procured for this present study, the activity profile of the wild type MtDapF was different from that reported in the authors' earlier paper (Usha et al., 2006). The apparent KM value for the wild-type M. tuberculosis DapF with meso-[3H]DAP was calculated at 166 μM, somewhat lower than the 1217 μM reported previously (Usha et al., 2006). Here, Kcat was determined as 0.1465 s−1 (Table 2), a turnover rate that is much lower than those described previously for DAP epimerases (Higgins et al., 1989; Koo & Blanchard, 1999). However, this is consistent with an expected kinetic isotope effect 3(V/K) ≈5.9 (Wiseman & Nichols, 1984) and the use of a mixture of 3H and 1H isotopes in these assays. The values for Kcat and Kcat/KM derived herein can only be used for comparative purposes within this study.
|Kcat (s−1)||KM (μM)||Kcat/KM (s−1 M−1)|
The analyses of the DAP epimerase activity of each of the various mutants revealed that substitution of either cysteine residue significantly compromises the activity of the enzyme (Table 2). Unlike the CysSer substitutions, the DAP epimerase activity of their alaninyl counterparts C87A and C226A was undetectable in vitro at all concentrations of meso-DAP and enzymes tested indicating that the two cysteine residues occupy critically important space that requires resident residues to accept and donate protons for catalysis. Consistently and like the glutamate racemase of L. fermenti (Glavas & Tanner, 1999), the CysSer substitutions retained residual activity (∼1% of Kcat) despite the substitution of the more stable alcohol (pKa∼16) for a thiol (pKa∼10) (Fig. 4).
The continuing global prevalence of tuberculosis and the alarming emergence of strains of the M. tuberculosis complex exhibiting multiple and extensive drug resistance profiles demand that new chemotherapeutic options are addressed. Historically, the biosynthesis of the mycobacterial cell wall has provided a useful chemotherapeutic target with component enzymes inhibited by isoniazid, ethionamide and ethambutol (Takayama & Kilburn, 1989; Banerjee et al., 1994; Kremer et al., 2003; Vilcheze et al., 2006). Despite the relatively unremarkable peptidoglycan structure, penicillins are not clinically effective against mycobacterial infections. Mycobacteria combine β-lactamase production (Chambers et al., 1995) with their ability to adopt an alternative, penicillin-insensitive stem peptide cross-linking mode in stationary phase, a strategy that is likely relevant in vivo (Goffin & Ghuysen, 2002). Other than the use of d-cycloserine, which inhibits stem peptide synthesis (Cáceres et al., 1997), as a second-line agent in drug-resistant cases (Ruiz, 1964), the inhibition of mycobacterial peptidoglycan has not been exploited clinically. The biosynthesis of meso-DAP represents a potentially useful node of mycobacterial cell wall physiology for chemical intervention as the di-amino acid is strictly required for all peptidoglycan stem peptide cross-linking (Wietzerbin et al., 1974). Thus, DAP epimerase was considered a good candidate for further investigation towards the development of new antimycobacterial agents. Recently, the authors overcame the particularly poor yields of recombinant DapF produced in E. coli by introducing several silent SW mutations in the 5′ end of the mycobacterial sequence and began to characterize the enzyme (Usha et al., 2006).
Here, the study was extended by confirming the likely mechanism of the epimerization reaction. The enzyme shows significant sequence similarity to its H. influenzae orthologue with 23.1 % identity. Of particular significance is the conservation of two cysteine residues that are catalytically active in other DAP epimerases as well as other PLP-independent amino acid racemases (Koo & Blanchard, 1999; Koo et al., 2000; Tanner, 2002). Consistent with their involvement here, treatment of MtDapF with nanomolar concentrations of three different thiol-specific alkylating agents resulted in the total ablation of DAP epimerase activity.
The authors were thus encouraged to construct substitution mutants at both of these residues; the wild-type Cys residues were replaced individually by both serine and alanine residues. The activities that were measured were completely consistent with the observations of Koo et al. (2000) using equivalent mutants of DapF from H. influenzae. In all cases, the epimerase activity of the mutant enzymes was severely impaired, the extent of the impairment being dependent on the nature of the amino acid replacing the Cys. Marked decreases in Kcat/KM revealed that replacement of either Cys87 or Cys226 with Ser produced mutant enzymes that were at least 50-fold less efficient than wild-type despite retaining similar or even improved substrate affinities indicated by the KM values exhibited for DAP.
As the catalytic efficiency of this mutant was severely diminished, the moderately improved KM for meso-DAP apparent with the C87S mutant indicates nonproductive substrate binding. Presumably here, the deprotonation of Cα and the formation of the planar carbanion intermediate is likely to be supported by the presence of Cys226, but its subsequent protonation by S87 is compromised. This engagement of the catalytic machinery may culminate in an increased residency of substrate in the active site without efficient epimerisation, thus explaining the decrease in the value of the apparent KM, i.e. the off-rate is slowed. Such a phenomenon is not apparent using the d,l-DAP substrate with C226S as the initial deprotonation is less favoured, although a similar effect might occur if l,l-DAP were used as the roles of the Cys residues are interchangeable.
In conclusion, all the observations are consistent with the hypothesis that MtDapF and DapF from H. influenzae achieve the epimerisation of DAP through an identical mechanism. Thus it can be confidently stated that the structural models produced are accurate enough to drive structure-based drug design.
The authors wish to thank Dr Dean Rea of the School of Biological Sciences, University of Warwick and Dr Sanjib Bhakta of the School of Biological and Chemical Sciences, Birkbeck, University of London for helpful discussions. V.U. acknowledges support from The Darwin Trust of Edinburgh. G.S.B. acknowledges the support of Mr James Bardrick in the form of a Personal Research Chair, and a Royal Society Wolfson Research Merit Award, as well as the Wellcome Trust and Medical Research council through various research grants.
- 1998) Properties of glutamate racemase from Bacillus subtilis IFO 3336 producing poly-gamma-glutamate. J Biochem (Tokyo) 123: 1156–1163. , , & (
- 1994) InhA, a gene encoding a target for isoniazid and ethionamide in Mycobacterium tuberculosis. Science 263: 227–230. , , et al. (
- 2006) BCG vaccine: efficacy and indications for vaccination and revaccination. J Pediatr (Rio J) 82: S45–S54. , & (
- 2007) Deletion of kasB in Mycobacterium tuberculosis causes loss of acid-fastness and subclinical latent tuberculosis in immunocompetent mice. Proc Natl Acad Sci USA 104: 5157–5162. , , et al. (
- 1999) Structure/function studies on enzymes in the diaminopimelate pathway of bacterial cell wall biosynthesis. Curr Opin Chem Biol 3: 607–613. & (
- 1997) Overexpression of the d-alanine racemase gene confers resistance to d-cycloserine in Mycobacterium smegmatis. J Bacteriol 179: 5046–5055. , , , , & (
- 1995) Can penicillins and other beta-lactam antibiotics be used to treat tuberculosis? Antimicrob Agents Chemother 39: 2620–2624. , , et al. (
- 1998) Structural symmetry: the three-dimensional structure of Haemophilus influenzae diaminopimelate epimerase. Biochemistry 37: 16452–16458. , , & (
- 2006) Global epidemiology of tuberculosis. Lancet 367: 938–940. (
- 1993) Purification, cloning, and cofactor independence of glutamate racemase from Lactobacillus. Biochemistry 32: 3981–3990. & (
- 2003) Requirement for kasB in Mycobacterium mycolic acid biosynthesis, cell wall impermeability and intracellular survival: implications for therapy. Mol Microbiol 49: 1547–1563. , , et al. (
- 1999) Catalytic acid/base residues of glutamate racemase. Biochemistry 38: 4106–4113. & (
- 2002) Biochemistry and comparative genomics of SxxK superfamily acyltransferases offer a clue to the mycobacterial paradox: presence of penicillin-susceptible target proteins versus lack of efficiency of penicillin as therapeutic agent. Microbiol Mol Biol Rev 66: 702–738. & (
- 2003) Identification and characterization of the last two unknown genes, dapC and dapF, in the succinylase branch of the l-lysine biosynthesis of Corynebacterium glutamicum. J Biotechnol 104: 199–211. , , , , , & (
- 1989) Expression of recombinant diaminopimelate epimerase in Escherichia coli. Isolation and inhibition with an irreversible inhibitor. Eur J Biochem 186: 137–143. , , , & (
- 2007) Changing funding patterns in tuberculosis. Nat Med 13: 299–303. & (
- 1999) Chemical mechanism of Haemophilus influenzae diaminopimelate epimerase. Biochemistry 38: 4416–4422. & (
- 2000) Identification of active site cysteine residues that function as general bases: diaminopimelate epimerase. J Am Chem Soc 122: 6122–6123. , , & (
- 2003) Inhibition of inhA activity, but not kasA activity, induces formation of a kasA-containing complex in mycobacteria. J Biol Chem 278: 20547–20554. , , et al. (
- 2004) Refinement of Haemophilus influenzae diaminopimelic acid epimerase (dapF) at 1.75 A resolution suggests a mechanism for stereocontrol during catalysis. Acta Crystallogr D Biol Crystallogr 60: 397–400. , , & (
- 2006) Structural insights into stereochemical inversion by diaminopimelate epimerase: an antibacterial drug target. Proc Natl Acad Sci USA 103: 8668–8673. , , , , , & (
- 1987) Molecular cloning, characterization, and chromosomal localization of dapF, the Escherichia coli gene for diaminopimelate epimerase. J Bacteriol 169: 1454–1459. , , & (
- 1975) Reaction mechanism and structure of the active site of proline racemase. Biochemistry 14: 4515–4522. & (
- 1964) d-cycloserine in the treatment of pulmonary tuberculosis resistant to the standard drugs: a study of 116 cases. Dis Chest 45: 181–186. (
- 1972) Peptidoglycan types of bacterial cell walls and their taxonomic implications. Bacteriol Rev 36: 407–477. & (
- 1989) Inhibition of synthesis of arabinogalactan by ethambutol in Mycobacterium smegmatis. Antimicrob Agents Chemother 33: 1493–1499. & (
- 2002) Understanding nature's strategies for enzyme-catalyzed racemization and epimerization. Acc Chem Res 35: 237–246. (
- 1993) Isotope effects and the identification of catalytic residues in the reaction catalyzed by glutamate racemase. Biochemistry 32: 3998–4006. , & (
- 2006) Use of a codon alteration strategy in a novel approach to cloning the Mycobacterium tuberculosis diaminopimelic acid epimerase. FEMS Microbiol Lett 262: 39–47. , , , & (
- 2006) Transfer of a point mutation in Mycobacterium tuberculosis inhA resolves the target of isoniazid. Nat Med 12: 1027–1029. , , et al. (
- 1974) Occurrence of d-alanyl-(d)-meso-diaminopimelic acid and meso-diaminopimelyl-meso-diaminopimelic acid interpeptide linkages in the peptidoglycan of Mycobacteria. Biochemistry 13: 3471–3476. , , , , & (
- 1970) Effects of isoniazid on mycolic acid synthesis in Mycobacterium tuberculosis and on its cell envelope. Biochem J 117: P27. , & (
- 1984) Purification and properties of diaminopimelic acid epimerase from Escherichia coli. J Biol Chem 259: 8907–8914. & (