Peptidoglycan is a major component of the bacterial cell wall and thus a major determinant of cell shape. Its biosynthesis is initiated by several sequential reactions catalyzed by cytoplasmic Mur enzymes. Mur ligases (MurC, -D, -E, and -F) are essential for bacteria, metabolize molecules not present in eukaryotes, and are structurally and biochemically tractable. However, although many Mur inhibitors have been developed, few have shown promising antibacterial activity, prompting the hypothesis that within the cytoplasm, Mur enzymes could exist as a complex whose architecture limits access of small molecules to their active sites. This suggestion is supported by the observation that in many bacteria, mur genes are present in a single operon, and pairs of these genes often are fused to generate a single polypeptide. Here, we explored this genetic arrangement in the human pathogen Bordetella pertussis and show that MurE and MurF are expressed as a single, bifunctional protein. EM, small angle X-ray scattering (SAXS), and analytical centrifugation (AUC) revealed that the MurE-MurF fusion displays an elongated, flexible structure that can dimerize. Moreover, MurE-MurF interacted with the peripheral glycosyltransferase MurG, which formed discrete oligomers resembling 4- or 5-armed stars in EM images. The oligomeric structure of MurG may allow it to play a bona fide scaffolding role for a potential Mur complex, facilitating the efficient conveyance of peptidoglycan-building blocks toward the inner membrane leaflet. Our findings shed light on the structural determinants of a peptidoglycan formation complex involving Mur enzymes in bacterial cell wall formation.
ABSTRACT Due to the rising incidence of antibiotic-resistant infections, the last-line antibiotics, polymyxins, have resurged in the clinics in parallel with new bacterial strategies of escape. The Gram-negative opportunistic pathogen Pseudomonas aeruginosa develops resistance to colistin/polymyxin B by distinct molecular mechanisms, mostly through modification of the lipid A component of the LPS by proteins encoded within the arnBCDATEF-ugd ( arn ) operon. In this work, we characterized a polymyxin-induced operon named mipBA , present in P. aeruginosa strains devoid of the arn operon. We showed that mipBA is activated by the ParR/ParS two-component regulatory system in response to polymyxins. Structural modeling revealed that MipA folds as an outer-membrane β-barrel, harboring an internal negatively charged channel, able to host a polymyxin molecule, while the lipoprotein MipB adopts a β-lactamase fold with two additional C-terminal domains. Experimental work confirmed that MipA and MipB localize to the bacterial envelope, and they co-purify in vitro . Nano differential scanning fluorimetry showed that polymyxins stabilized MipA in a specific and dose-dependent manner. Mass spectrometry-based quantitative proteomics on P. aeruginosa membranes demonstrated that ∆ mipBA synthesized fourfold less MexXY-OprA proteins in response to polymyxin B compared to the wild-type strain. The decrease was a direct consequence of impaired transcriptional activation of the mex operon operated by ParR/ParS. We propose MipA/MipB to act as membrane (co)sensors working in concert to activate ParS histidine kinase and help the bacterium to cope with polymyxin-mediated envelope stress through synthesis of the efflux pump, MexXY-OprA. IMPORTANCE Due to the emergence of multidrug-resistant isolates, antibiotic options may be limited to polymyxins to eradicate Gram-negative infections. Pseudomonas aeruginosa , a leading opportunistic pathogen, has the ability to develop resistance to these cationic lipopeptides by modifying its lipopolysaccharide through proteins encoded within the arn operon. Herein, we describe a sub-group of P. aeruginosa strains lacking the arn operon yet exhibiting adaptability to polymyxins. Exposition to sub-lethal polymyxin concentrations induced the expression and production of two envelope-associated proteins. Among those, MipA, an outer-membrane barrel, is able to specifically bind polymyxins with an affinity in the 10-µM range. Using membrane proteomics and phenotypic assays, we showed that MipA and MipB participate in the adaptive response to polymyxins via ParR/ParS regulatory signaling. We propose a new model wherein the MipA-MipB module functions as a novel polymyxin sensing mechanism.
Type III secretion (T3S) systems allow the export and translocation of bacterial effectors into the host cell cytoplasm. Secretion is accomplished by an 80-nm-long needle-like structure composed, in Pseudomonas aeruginosa, of the polymerized form of a 7-kDa protein, PscF. Two proteins, PscG and PscE, stabilize PscF within the bacterial cell before its export and polymerization. In this work we screened the 1,320-A(2) interface between the two chaperones, PscE and PscG, by site-directed mutagenesis and determined hot spot regions that are important for T3S function in vivo and complex formation in vitro. Three amino acids in PscE and five amino acids in PscG, found to be relevant for complex formation, map to the central part of the interacting surface. Stability assays on selected mutants performed both in vitro on purified PscE-PscG complexes and in vivo on P. aeruginosa revealed that PscE is a cochaperone that is essential for the stability of the main chaperone, PscG. Notably, when overexpressed from a bicistronic construct, PscG and PscF compensate for the absence of PscE in cytotoxic P. aeruginosa. These results show that all of the information needed for needle protein stabilization and folding, its presentation to the T3 secreton, and its export is present within the sequence of the PscG chaperone.
Arg285, one of the very few conserved residues in the active site of d-amino acid oxidases, has been mutated to lysine, glutamine, aspartate, and alanine in the enzyme from the yeast Rhodotorula gracilis (RgDAAO). The mutated proteins are all catalytically competent. Mutations of Arg285 result in an increase (≈300-fold) ofK m for the d-amino acid and in a large decrease (≈500-fold) of turnover number. Stopped-flow analysis shows that the decrease in turnover is paralleled by a similar decrease in the rate of flavin reduction (k 2), the latter still being the rate-limiting step of the reaction. In agreement with data from the protein crystal structure, loss of the guanidinium group of Arg285 in the mutated DAAOs drastically reduces the binding of several carboxylic acids (e.g. benzoate). These results highlight the importance of this active site residue in the precise substrate orientation, a main factor in this redox reaction. Furthermore, Arg285 DAAO mutants have spectral properties similar to those of the wild-type enzyme, but show a low degree of stabilization of the flavin semiquinone and a change in the redox properties of the free enzyme. From this, we can unexpectedly conclude that Arg285 in the free enzyme form is involved in the stabilization of the negative charge on the N(1)-C(2)=O locus of the isoalloxazine ring of the flavin. We also suggest that the residue undergoes a conformational change in order to bind the carboxylate portion of the substrate/ligand in the complexed enzyme. Arg285, one of the very few conserved residues in the active site of d-amino acid oxidases, has been mutated to lysine, glutamine, aspartate, and alanine in the enzyme from the yeast Rhodotorula gracilis (RgDAAO). The mutated proteins are all catalytically competent. Mutations of Arg285 result in an increase (≈300-fold) ofK m for the d-amino acid and in a large decrease (≈500-fold) of turnover number. Stopped-flow analysis shows that the decrease in turnover is paralleled by a similar decrease in the rate of flavin reduction (k 2), the latter still being the rate-limiting step of the reaction. In agreement with data from the protein crystal structure, loss of the guanidinium group of Arg285 in the mutated DAAOs drastically reduces the binding of several carboxylic acids (e.g. benzoate). These results highlight the importance of this active site residue in the precise substrate orientation, a main factor in this redox reaction. Furthermore, Arg285 DAAO mutants have spectral properties similar to those of the wild-type enzyme, but show a low degree of stabilization of the flavin semiquinone and a change in the redox properties of the free enzyme. From this, we can unexpectedly conclude that Arg285 in the free enzyme form is involved in the stabilization of the negative charge on the N(1)-C(2)=O locus of the isoalloxazine ring of the flavin. We also suggest that the residue undergoes a conformational change in order to bind the carboxylate portion of the substrate/ligand in the complexed enzyme. d-amino acid oxidase (EC 1.4.3.3) Rhodotorula gracilis d-amino acid oxidase pig kidney d-amino acid oxidase oxidized enzyme enzyme flavin semiquinone reduced enzyme 3,3,3-trifluoro-d-alanine d-Amino acid oxidase (EC 1.4.3.3, DAAO)1 catalyzes the dehydrogenation of d-isomer of amino acids to give the corresponding α-imino acids and, after subsequent hydrolysis, α-keto acids and ammonia. New interesting findings (as a role of the enzyme in modulation d-serine level in brain) shed light on the debated role of DAAO in mammalian organisms (1Schell M.J. Molliver M.E. Snyder S.H. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3948-3952Crossref PubMed Scopus (745) Google Scholar, 2Wolosker H. Blackshaw S. Snyder S.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 13409-13414Crossref PubMed Scopus (711) Google Scholar). The precise mechanism of substrate dehydrogenation of this well studied enzyme (3Curti B. Ronchi S. Pilone Simonetta M. Muller F. Chemistry and Biochemistry of Flavoenzymes. CRC Press, Boca Raton, FL1992: 69-94Google Scholar) has not yet been solved, even if recently two groups have reported the crystal structure of the enzyme purified from pig kidney (pkDAAO) at a resolution of 2.6 and 3.0 Å, respectively (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar). Over the years, three main different mechanisms have been proposed for the reaction catalyzed by this flavoenzyme (see Mattevi et al. (6Mattevi A. Vanoni M.A. Curti B. Curr. Opin. Struct. Biol. 1997; 7: 804-810Crossref PubMed Scopus (31) Google Scholar) for a recent review). (i) The hypothesis that the reductive half-reaction of DAAO involves the initial formation of a carbanion by abstraction of the α-H of the substrate as a proton comes from the elimination of halide from β-chloro-d-alanine (7Walsh C.T. Schonbrunn A. Abeles R. J. Biol. Chem. 1971; 246: 6855-6866Abstract Full Text PDF PubMed Google Scholar). (ii) The observation of a transfer of α-hydrogen of the substrate to the C(5) position of the enzyme reconstituted with 5-deaza-FAD provides evidence in favor of a direct hydride-transfer mechanism (8Hersh L.B. Jorns M.S. J. Biol. Chem. 1975; 250: 8728-8735Abstract Full Text PDF PubMed Google Scholar). (iii) Finally a concerted mechanism (consistent with the experimental evidence for a carbanion mechanism) in which α-H+ abstraction is coupled with the transfer of a hydride from the amino group of the substrate has been put forward (9Miura R. Miyake Y. Bioorg. Chem. 1988; 16: 97-110Crossref Scopus (30) Google Scholar).As a model for DAAO to have a better understanding of this crucial issue, we have used the enzyme from the red yeast Rhodotorula gracilis (RgDAAO). Actually, RgDAAO possesses peculiar properties, as the high catalytic efficiency and the tight binding with the coenzyme FAD, which distinguish it from the mammalian enzyme (10Pilone Simonetta M. Pollegioni L. Casalin P. Curti B. Ronchi S. Eur. J. Biochem. 1989; 180: 199-204Crossref PubMed Scopus (63) Google Scholar, 11Casalin P. Pollegioni L. Curti B. Pilone M.S. Eur. J. Biochem. 1991; 197: 513-517Crossref PubMed Scopus (61) Google Scholar, 12Pollegioni L. Langkau B. Tischer W. Ghisla S. Pilone M.S. J. Biol. Chem. 1993; 268: 13850-13857Abstract Full Text PDF PubMed Google Scholar). These properties are most probably related to its physiological role (yeast can metabolize d-amino acids and use them as the sole nitrogen and carbon source) and to an evolutionary drive. From comparison of the primary sequences of the known DAAOs, it is evident that only three residues, among those identified in or near the active site (13Faotto L. Pollegioni L. Ceciliani F. Ronchi S. Pilone M.S. Biotechnol. Lett. 1995; 17: 193-198Crossref Scopus (47) Google Scholar), are conserved (namely two tyrosines and one arginine). The presence of an arginine residue located at the active site and directly involved in DAAO catalysis was in fact previously proposed by various chemical modification studies (for a review, see Ref. 3Curti B. Ronchi S. Pilone Simonetta M. Muller F. Chemistry and Biochemistry of Flavoenzymes. CRC Press, Boca Raton, FL1992: 69-94Google Scholar and references therein). The possibility that the arginine residue could be involved in substrate binding by electrostatic interaction with the substrate was inferred by Nishino et al. (14Nishino T. Massey V. Williams Jr., C.H. J. Biol. Chem. 1980; 255: 3610-3616Abstract Full Text PDF PubMed Google Scholar) from 2,3-butanedione modification of the mammalian enzyme, followed by reaction with dansylchloride. On the other hand, the reactivity with sulfite and the spectral properties of the native and cyclohexandione-modified pkDAAO reconstituted with 8-mercapto-FAD suggested that the active site arginine could act as the positively charged group near the flavin N(1)-C(2)=O locus and responsible for stabilization of anionic flavin forms (15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar). This basic residue has been identified in the primary sequence of RgDAAO, by irreversible inhibition using phenylglyoxal (16Gadda G. Negri A. Pilone M.S. J. Biol. Chem. 1994; 269: 17809-17814Abstract Full Text PDF PubMed Google Scholar), although the question regarding its role was not solved. More recently, the resolution of the crystal structure of pkDAAO in complex with benzoate and anthranilate (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar) showed that Arg283interacts with the carboxylate group of the ligand and that the negative charge on the N(1)-C(2)=O locus of the flavin is stabilized by the dipole of α-helix F5. This result was apparently in contrast with the previous reports (see above).The crystal structure of RgDAAO has been recently solved at very high resolution (up to 1.2 Å), as a basis for the interpretation of the mechanistic studies and to find a rationale of its high catalytic efficiency. 2S. Umhau, L. Pollegioni, G. Molla, K. Diederichs, W. Welte, M. S. Pilone, and S. Ghisla, submitted for publication.2S. Umhau, L. Pollegioni, G. Molla, K. Diederichs, W. Welte, M. S. Pilone, and S. Ghisla, submitted for publication. The structure of oxidized enzyme in complex with the quasi-substrate CF3-alanine and of reduced enzyme in complex with the substrate d-alanine, revealed the mode of substrate binding (Fig. 1 A). The α-carboxylic group of the d-amino acid interacts electrostatically with the γ- and ε-amino groups of Arg285 (at ∼2.8 Å) and it is H-bonded with the hydroxyl groups of Tyr223 and Tyr238. The substrate α-amino group is H-bonded symmetrically with the backbone C=O group of Ser335 and the active site water molecule H2O72, while the substrate side chain is oriented toward the hydrophobic binding pocket of the active site (see Fig. 1 A).At the same time, we are substantiating the role of the active site residues of RgDAAO by site-directed mutagenesis of each single residue. In a previous paper we reported the effect of substitution of Tyr223 with a phenylalanine and a serine (18Harris C.M. Molla G. Pilone M.S. Pollegioni L. J. Biol. Chem. 1999; 274: 36233-36240Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). The characterization of the corresponding mutant RgDAAO's allowed us to exclude that Tyr223 can act as an active-site base, and highlighted its importance for the correct orientation of the bound substrate, as required for an efficient catalysis. We report here on the characterization of four single points mutants obtained by substitution of Arg285. The combination of the information derived from the site-directed mutagenesis studies and from the crystal structures confirms the involvement of Arg285 in substrate binding, but put forward a possible role in stabilization of the negative charge on the flavin N(1)-C(2)=O locus in the free enzyme.DISCUSSIONThe results presented in this paper demonstrate that the conserved Arg285 plays different roles at the active site of RgDAAO. We successfully expressed in E. coli four RgDAAO mutants at position Arg285 using the pT7-DAAO expression system (21Molla G. Vegezzi C. Pilone M.S. Pollegioni L. Protein Expression Purif. 1998; 14: 289-294Crossref PubMed Scopus (65) Google Scholar) and purified them to homogeneity as holoenzyme with a good yield (≥20%). The mutation of Arg285 results in no gross perturbation or loss of FAD, thus the observed changes are due to only specific and local structural modifications. The main role of this residue in substrate binding has been substantiated by the known three-dimensional structure of mammalian DAAO (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar) and, recently, from the structure at high resolution of RgDAAO.2 In the latter, the structure of the reduced enzyme-d-alanine complex at 1.2-Å resolution reveals the mode of substrate binding (Fig. 1 A): the carboxylate position of the amino acid is electrostatically bound via a two-point interaction with the γ- and ε-amino group of Arg285 (at a distance of 2.79 Å).2 The large decrease in substrate affinity (and ligand binding) observed with the mutant enzymes is well explained by the involvement of this arginine residue in binding and fixation (Fig.1 A). The ligand-binding experiments demonstrate that the overall substrate binding pocket is largely altered even when a conservative mutation, as in the case of the R285K mutant, is present (Table III).The change in flavin redox potentials of R285K and the thermodynamic instability of the semiquinone form in all mutants profoundly distinguish them from the wild-type DAAO. At the moment, only the three-dimensional structure of RgDAAO and pkDAAO in complex with a substrate or a ligand is available (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar).2 In this structure, the isoalloxazine ring is located in a hydrophobic environment making contacts with the side chain of different residues but, differently from pkDAAO and from other flavoproteins (6Mattevi A. Vanoni M.A. Curti B. Curr. Opin. Struct. Biol. 1997; 7: 804-810Crossref PubMed Scopus (31) Google Scholar), no basic residue(s) or α-helix dipole is properly located to interact with the flavin N(1)-C(2)=O position.2 A (partial) positive charge in proximity of this flavin locus is required to stabilize several anionic flavin derivatives (e.g. the N(5) covalent adduct with sulfite). Since the ε-amino group of Arg285 in RgDAAO is ∼7 Å far away from the N(1)-C(2)=O flavin position (in the EFlred-d-alanine complex), it cannot stabilize the negative charge on the flavin.2 We suggest that in the free enzyme form (i.e. in the absence of a ligand) the side chain of Arg285 is able to rotate to a distance of ∼3 Å from the N(1)-C(2)=O flavin locus (Fig. 1 B). Therefore, the lack of the guanidinium group of Arg285 in the mutants determines the absence of the counterion required for stabilization of flavin semiquinone. This lower stabilization is further supported by the ∼104-fold increase in K d for binding of sulfite (Table III). Moreover, we are tempted to conclude that a similar function is exerted by Arg283 in the mammalian DAAO, and thus to give a rationale of the results obtained from chemical modification studies (15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar, 34Ferti C. Curti B. Simonetta M.S. Ronchi S. Galliano M. Minchiotti L. Eur. J. Biochem. 1981; 119: 553-557Crossref PubMed Scopus (20) Google Scholar). In fact, the inferred interaction of Arg283 with the flavin N(1)-C(2)=O locus in the free form of pkDAAO could explain the observation that the chemical modification of this residue destroyed the ability of pkDAAO to stabilize the benzoquinoid form of 8-mercapto-FAD and to form an N(5)-adduct with sulfite (15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar).The crystal structure of oxidized DAAO in complex withl-lactate and of reduced enzyme complexed withd-alanine shows that the side chain of Arg285is at >5 Å from the α-CH group of the substrate. Thereby, in this position Arg285 is far way from the reactive α-hydrogen and cannot be the active site base required by a carbanion mechanism for substrate dehydrogenation. In fact, all the Arg285mutants we produced possess appreciable dehydrogenase activity (they can be anaerobically reduced by the substrate d-alanine). On the other hand, the large decrease in the rate of flavin reduction (and therefore in turnover) is quite surprising. Thek cat for R285K and R285A are decreased by about 450- and 7000-fold, respectively, in comparison to the wild-type RgDAAO (12Pollegioni L. Langkau B. Tischer W. Ghisla S. Pilone M.S. J. Biol. Chem. 1993; 268: 13850-13857Abstract Full Text PDF PubMed Google Scholar); this change is accompanied by a parallel decrease in the rate of flavin reduction (Table IV). Since k 2 is still rate-limiting, we can rule out a change in the kinetic mechanism. A main kinetic difference with respect to the wild-type DAAO is that, due to the large decrease in k 2,k 3 ≫ k 2 for the Arg285 mutants and thus the reductive half-reaction is essentially monophasic. The effect of Arg285 substitution on RgDAAO catalysis can be explained in terms of the recently proposed mechanism in which "orbital steering/interactions are the predominant or the sole important factors in catalysis."2 The perturbation of the active site in the Arg285 mutants modifies the precise substrate alignment: alteration of the reaction trajectory results in a large change in the reaction velocity (k red and k cat). A similar large effect has been also reported for the enzyme isocitrate dehydrogenase, as a consequence of small trajectory changes via substrate modification and metal co-ordination (17Mesecar A.D. Stoddard B.L. Koshland D.E. Science. 1997; 277: 202-206Crossref PubMed Scopus (200) Google Scholar). Actually, a second possibility has also to be taken into account, namely that an equilibrium form of the enzyme-substrate complex exists, in which the guanidinium side chain of Arg285 is no longer in contact with the carboxylate but it reaches a position in which it could abstract the proton from the α-carbon of the substrate.The mechanism by which substrate dehydrogenation occurs in DAAO cannot be solved solely on the basis of structural data, thus new investigations are required to rule out definitively the possibility that Arg285 could be the acid/base residue required by a carbanion-type mechanism. Our results support the concept that precise binding and orientation of the substrate is a main quantitative factor in RgDAAO catalysis. Albeit the lack of the free enzyme crystal structure, the novel hypothesis on a conformational swing of the side chain of Arg285 from the position (seen in the structure) in which it binds the carboxylate anion, to a different one (next to the flavin N(1)-C(2)=O locus) in the free enzyme form, fits with the interpretation of the whole body of data. Moreover, the model we have proposed permits a reconciliation of the apparently contradictory conclusions that arose from the three-dimensional structure and from chemical modification studies of pkDAAO (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar, 14Nishino T. Massey V. Williams Jr., C.H. J. Biol. Chem. 1980; 255: 3610-3616Abstract Full Text PDF PubMed Google Scholar, 15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar). d-Amino acid oxidase (EC 1.4.3.3, DAAO)1 catalyzes the dehydrogenation of d-isomer of amino acids to give the corresponding α-imino acids and, after subsequent hydrolysis, α-keto acids and ammonia. New interesting findings (as a role of the enzyme in modulation d-serine level in brain) shed light on the debated role of DAAO in mammalian organisms (1Schell M.J. Molliver M.E. Snyder S.H. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3948-3952Crossref PubMed Scopus (745) Google Scholar, 2Wolosker H. Blackshaw S. Snyder S.H. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 13409-13414Crossref PubMed Scopus (711) Google Scholar). The precise mechanism of substrate dehydrogenation of this well studied enzyme (3Curti B. Ronchi S. Pilone Simonetta M. Muller F. Chemistry and Biochemistry of Flavoenzymes. CRC Press, Boca Raton, FL1992: 69-94Google Scholar) has not yet been solved, even if recently two groups have reported the crystal structure of the enzyme purified from pig kidney (pkDAAO) at a resolution of 2.6 and 3.0 Å, respectively (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar). Over the years, three main different mechanisms have been proposed for the reaction catalyzed by this flavoenzyme (see Mattevi et al. (6Mattevi A. Vanoni M.A. Curti B. Curr. Opin. Struct. Biol. 1997; 7: 804-810Crossref PubMed Scopus (31) Google Scholar) for a recent review). (i) The hypothesis that the reductive half-reaction of DAAO involves the initial formation of a carbanion by abstraction of the α-H of the substrate as a proton comes from the elimination of halide from β-chloro-d-alanine (7Walsh C.T. Schonbrunn A. Abeles R. J. Biol. Chem. 1971; 246: 6855-6866Abstract Full Text PDF PubMed Google Scholar). (ii) The observation of a transfer of α-hydrogen of the substrate to the C(5) position of the enzyme reconstituted with 5-deaza-FAD provides evidence in favor of a direct hydride-transfer mechanism (8Hersh L.B. Jorns M.S. J. Biol. Chem. 1975; 250: 8728-8735Abstract Full Text PDF PubMed Google Scholar). (iii) Finally a concerted mechanism (consistent with the experimental evidence for a carbanion mechanism) in which α-H+ abstraction is coupled with the transfer of a hydride from the amino group of the substrate has been put forward (9Miura R. Miyake Y. Bioorg. Chem. 1988; 16: 97-110Crossref Scopus (30) Google Scholar). As a model for DAAO to have a better understanding of this crucial issue, we have used the enzyme from the red yeast Rhodotorula gracilis (RgDAAO). Actually, RgDAAO possesses peculiar properties, as the high catalytic efficiency and the tight binding with the coenzyme FAD, which distinguish it from the mammalian enzyme (10Pilone Simonetta M. Pollegioni L. Casalin P. Curti B. Ronchi S. Eur. J. Biochem. 1989; 180: 199-204Crossref PubMed Scopus (63) Google Scholar, 11Casalin P. Pollegioni L. Curti B. Pilone M.S. Eur. J. Biochem. 1991; 197: 513-517Crossref PubMed Scopus (61) Google Scholar, 12Pollegioni L. Langkau B. Tischer W. Ghisla S. Pilone M.S. J. Biol. Chem. 1993; 268: 13850-13857Abstract Full Text PDF PubMed Google Scholar). These properties are most probably related to its physiological role (yeast can metabolize d-amino acids and use them as the sole nitrogen and carbon source) and to an evolutionary drive. From comparison of the primary sequences of the known DAAOs, it is evident that only three residues, among those identified in or near the active site (13Faotto L. Pollegioni L. Ceciliani F. Ronchi S. Pilone M.S. Biotechnol. Lett. 1995; 17: 193-198Crossref Scopus (47) Google Scholar), are conserved (namely two tyrosines and one arginine). The presence of an arginine residue located at the active site and directly involved in DAAO catalysis was in fact previously proposed by various chemical modification studies (for a review, see Ref. 3Curti B. Ronchi S. Pilone Simonetta M. Muller F. Chemistry and Biochemistry of Flavoenzymes. CRC Press, Boca Raton, FL1992: 69-94Google Scholar and references therein). The possibility that the arginine residue could be involved in substrate binding by electrostatic interaction with the substrate was inferred by Nishino et al. (14Nishino T. Massey V. Williams Jr., C.H. J. Biol. Chem. 1980; 255: 3610-3616Abstract Full Text PDF PubMed Google Scholar) from 2,3-butanedione modification of the mammalian enzyme, followed by reaction with dansylchloride. On the other hand, the reactivity with sulfite and the spectral properties of the native and cyclohexandione-modified pkDAAO reconstituted with 8-mercapto-FAD suggested that the active site arginine could act as the positively charged group near the flavin N(1)-C(2)=O locus and responsible for stabilization of anionic flavin forms (15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar). This basic residue has been identified in the primary sequence of RgDAAO, by irreversible inhibition using phenylglyoxal (16Gadda G. Negri A. Pilone M.S. J. Biol. Chem. 1994; 269: 17809-17814Abstract Full Text PDF PubMed Google Scholar), although the question regarding its role was not solved. More recently, the resolution of the crystal structure of pkDAAO in complex with benzoate and anthranilate (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar) showed that Arg283interacts with the carboxylate group of the ligand and that the negative charge on the N(1)-C(2)=O locus of the flavin is stabilized by the dipole of α-helix F5. This result was apparently in contrast with the previous reports (see above). The crystal structure of RgDAAO has been recently solved at very high resolution (up to 1.2 Å), as a basis for the interpretation of the mechanistic studies and to find a rationale of its high catalytic efficiency. 2S. Umhau, L. Pollegioni, G. Molla, K. Diederichs, W. Welte, M. S. Pilone, and S. Ghisla, submitted for publication.2S. Umhau, L. Pollegioni, G. Molla, K. Diederichs, W. Welte, M. S. Pilone, and S. Ghisla, submitted for publication. The structure of oxidized enzyme in complex with the quasi-substrate CF3-alanine and of reduced enzyme in complex with the substrate d-alanine, revealed the mode of substrate binding (Fig. 1 A). The α-carboxylic group of the d-amino acid interacts electrostatically with the γ- and ε-amino groups of Arg285 (at ∼2.8 Å) and it is H-bonded with the hydroxyl groups of Tyr223 and Tyr238. The substrate α-amino group is H-bonded symmetrically with the backbone C=O group of Ser335 and the active site water molecule H2O72, while the substrate side chain is oriented toward the hydrophobic binding pocket of the active site (see Fig. 1 A). At the same time, we are substantiating the role of the active site residues of RgDAAO by site-directed mutagenesis of each single residue. In a previous paper we reported the effect of substitution of Tyr223 with a phenylalanine and a serine (18Harris C.M. Molla G. Pilone M.S. Pollegioni L. J. Biol. Chem. 1999; 274: 36233-36240Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). The characterization of the corresponding mutant RgDAAO's allowed us to exclude that Tyr223 can act as an active-site base, and highlighted its importance for the correct orientation of the bound substrate, as required for an efficient catalysis. We report here on the characterization of four single points mutants obtained by substitution of Arg285. The combination of the information derived from the site-directed mutagenesis studies and from the crystal structures confirms the involvement of Arg285 in substrate binding, but put forward a possible role in stabilization of the negative charge on the flavin N(1)-C(2)=O locus in the free enzyme. DISCUSSIONThe results presented in this paper demonstrate that the conserved Arg285 plays different roles at the active site of RgDAAO. We successfully expressed in E. coli four RgDAAO mutants at position Arg285 using the pT7-DAAO expression system (21Molla G. Vegezzi C. Pilone M.S. Pollegioni L. Protein Expression Purif. 1998; 14: 289-294Crossref PubMed Scopus (65) Google Scholar) and purified them to homogeneity as holoenzyme with a good yield (≥20%). The mutation of Arg285 results in no gross perturbation or loss of FAD, thus the observed changes are due to only specific and local structural modifications. The main role of this residue in substrate binding has been substantiated by the known three-dimensional structure of mammalian DAAO (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar) and, recently, from the structure at high resolution of RgDAAO.2 In the latter, the structure of the reduced enzyme-d-alanine complex at 1.2-Å resolution reveals the mode of substrate binding (Fig. 1 A): the carboxylate position of the amino acid is electrostatically bound via a two-point interaction with the γ- and ε-amino group of Arg285 (at a distance of 2.79 Å).2 The large decrease in substrate affinity (and ligand binding) observed with the mutant enzymes is well explained by the involvement of this arginine residue in binding and fixation (Fig.1 A). The ligand-binding experiments demonstrate that the overall substrate binding pocket is largely altered even when a conservative mutation, as in the case of the R285K mutant, is present (Table III).The change in flavin redox potentials of R285K and the thermodynamic instability of the semiquinone form in all mutants profoundly distinguish them from the wild-type DAAO. At the moment, only the three-dimensional structure of RgDAAO and pkDAAO in complex with a substrate or a ligand is available (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar).2 In this structure, the isoalloxazine ring is located in a hydrophobic environment making contacts with the side chain of different residues but, differently from pkDAAO and from other flavoproteins (6Mattevi A. Vanoni M.A. Curti B. Curr. Opin. Struct. Biol. 1997; 7: 804-810Crossref PubMed Scopus (31) Google Scholar), no basic residue(s) or α-helix dipole is properly located to interact with the flavin N(1)-C(2)=O position.2 A (partial) positive charge in proximity of this flavin locus is required to stabilize several anionic flavin derivatives (e.g. the N(5) covalent adduct with sulfite). Since the ε-amino group of Arg285 in RgDAAO is ∼7 Å far away from the N(1)-C(2)=O flavin position (in the EFlred-d-alanine complex), it cannot stabilize the negative charge on the flavin.2 We suggest that in the free enzyme form (i.e. in the absence of a ligand) the side chain of Arg285 is able to rotate to a distance of ∼3 Å from the N(1)-C(2)=O flavin locus (Fig. 1 B). Therefore, the lack of the guanidinium group of Arg285 in the mutants determines the absence of the counterion required for stabilization of flavin semiquinone. This lower stabilization is further supported by the ∼104-fold increase in K d for binding of sulfite (Table III). Moreover, we are tempted to conclude that a similar function is exerted by Arg283 in the mammalian DAAO, and thus to give a rationale of the results obtained from chemical modification studies (15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar, 34Ferti C. Curti B. Simonetta M.S. Ronchi S. Galliano M. Minchiotti L. Eur. J. Biochem. 1981; 119: 553-557Crossref PubMed Scopus (20) Google Scholar). In fact, the inferred interaction of Arg283 with the flavin N(1)-C(2)=O locus in the free form of pkDAAO could explain the observation that the chemical modification of this residue destroyed the ability of pkDAAO to stabilize the benzoquinoid form of 8-mercapto-FAD and to form an N(5)-adduct with sulfite (15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar).The crystal structure of oxidized DAAO in complex withl-lactate and of reduced enzyme complexed withd-alanine shows that the side chain of Arg285is at >5 Å from the α-CH group of the substrate. Thereby, in this position Arg285 is far way from the reactive α-hydrogen and cannot be the active site base required by a carbanion mechanism for substrate dehydrogenation. In fact, all the Arg285mutants we produced possess appreciable dehydrogenase activity (they can be anaerobically reduced by the substrate d-alanine). On the other hand, the large decrease in the rate of flavin reduction (and therefore in turnover) is quite surprising. Thek cat for R285K and R285A are decreased by about 450- and 7000-fold, respectively, in comparison to the wild-type RgDAAO (12Pollegioni L. Langkau B. Tischer W. Ghisla S. Pilone M.S. J. Biol. Chem. 1993; 268: 13850-13857Abstract Full Text PDF PubMed Google Scholar); this change is accompanied by a parallel decrease in the rate of flavin reduction (Table IV). Since k 2 is still rate-limiting, we can rule out a change in the kinetic mechanism. A main kinetic difference with respect to the wild-type DAAO is that, due to the large decrease in k 2,k 3 ≫ k 2 for the Arg285 mutants and thus the reductive half-reaction is essentially monophasic. The effect of Arg285 substitution on RgDAAO catalysis can be explained in terms of the recently proposed mechanism in which "orbital steering/interactions are the predominant or the sole important factors in catalysis."2 The perturbation of the active site in the Arg285 mutants modifies the precise substrate alignment: alteration of the reaction trajectory results in a large change in the reaction velocity (k red and k cat). A similar large effect has been also reported for the enzyme isocitrate dehydrogenase, as a consequence of small trajectory changes via substrate modification and metal co-ordination (17Mesecar A.D. Stoddard B.L. Koshland D.E. Science. 1997; 277: 202-206Crossref PubMed Scopus (200) Google Scholar). Actually, a second possibility has also to be taken into account, namely that an equilibrium form of the enzyme-substrate complex exists, in which the guanidinium side chain of Arg285 is no longer in contact with the carboxylate but it reaches a position in which it could abstract the proton from the α-carbon of the substrate.The mechanism by which substrate dehydrogenation occurs in DAAO cannot be solved solely on the basis of structural data, thus new investigations are required to rule out definitively the possibility that Arg285 could be the acid/base residue required by a carbanion-type mechanism. Our results support the concept that precise binding and orientation of the substrate is a main quantitative factor in RgDAAO catalysis. Albeit the lack of the free enzyme crystal structure, the novel hypothesis on a conformational swing of the side chain of Arg285 from the position (seen in the structure) in which it binds the carboxylate anion, to a different one (next to the flavin N(1)-C(2)=O locus) in the free enzyme form, fits with the interpretation of the whole body of data. Moreover, the model we have proposed permits a reconciliation of the apparently contradictory conclusions that arose from the three-dimensional structure and from chemical modification studies of pkDAAO (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar, 14Nishino T. Massey V. Williams Jr., C.H. J. Biol. Chem. 1980; 255: 3610-3616Abstract Full Text PDF PubMed Google Scholar, 15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar). The results presented in this paper demonstrate that the conserved Arg285 plays different roles at the active site of RgDAAO. We successfully expressed in E. coli four RgDAAO mutants at position Arg285 using the pT7-DAAO expression system (21Molla G. Vegezzi C. Pilone M.S. Pollegioni L. Protein Expression Purif. 1998; 14: 289-294Crossref PubMed Scopus (65) Google Scholar) and purified them to homogeneity as holoenzyme with a good yield (≥20%). The mutation of Arg285 results in no gross perturbation or loss of FAD, thus the observed changes are due to only specific and local structural modifications. The main role of this residue in substrate binding has been substantiated by the known three-dimensional structure of mammalian DAAO (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar) and, recently, from the structure at high resolution of RgDAAO.2 In the latter, the structure of the reduced enzyme-d-alanine complex at 1.2-Å resolution reveals the mode of substrate binding (Fig. 1 A): the carboxylate position of the amino acid is electrostatically bound via a two-point interaction with the γ- and ε-amino group of Arg285 (at a distance of 2.79 Å).2 The large decrease in substrate affinity (and ligand binding) observed with the mutant enzymes is well explained by the involvement of this arginine residue in binding and fixation (Fig.1 A). The ligand-binding experiments demonstrate that the overall substrate binding pocket is largely altered even when a conservative mutation, as in the case of the R285K mutant, is present (Table III). The change in flavin redox potentials of R285K and the thermodynamic instability of the semiquinone form in all mutants profoundly distinguish them from the wild-type DAAO. At the moment, only the three-dimensional structure of RgDAAO and pkDAAO in complex with a substrate or a ligand is available (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar).2 In this structure, the isoalloxazine ring is located in a hydrophobic environment making contacts with the side chain of different residues but, differently from pkDAAO and from other flavoproteins (6Mattevi A. Vanoni M.A. Curti B. Curr. Opin. Struct. Biol. 1997; 7: 804-810Crossref PubMed Scopus (31) Google Scholar), no basic residue(s) or α-helix dipole is properly located to interact with the flavin N(1)-C(2)=O position.2 A (partial) positive charge in proximity of this flavin locus is required to stabilize several anionic flavin derivatives (e.g. the N(5) covalent adduct with sulfite). Since the ε-amino group of Arg285 in RgDAAO is ∼7 Å far away from the N(1)-C(2)=O flavin position (in the EFlred-d-alanine complex), it cannot stabilize the negative charge on the flavin.2 We suggest that in the free enzyme form (i.e. in the absence of a ligand) the side chain of Arg285 is able to rotate to a distance of ∼3 Å from the N(1)-C(2)=O flavin locus (Fig. 1 B). Therefore, the lack of the guanidinium group of Arg285 in the mutants determines the absence of the counterion required for stabilization of flavin semiquinone. This lower stabilization is further supported by the ∼104-fold increase in K d for binding of sulfite (Table III). Moreover, we are tempted to conclude that a similar function is exerted by Arg283 in the mammalian DAAO, and thus to give a rationale of the results obtained from chemical modification studies (15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar, 34Ferti C. Curti B. Simonetta M.S. Ronchi S. Galliano M. Minchiotti L. Eur. J. Biochem. 1981; 119: 553-557Crossref PubMed Scopus (20) Google Scholar). In fact, the inferred interaction of Arg283 with the flavin N(1)-C(2)=O locus in the free form of pkDAAO could explain the observation that the chemical modification of this residue destroyed the ability of pkDAAO to stabilize the benzoquinoid form of 8-mercapto-FAD and to form an N(5)-adduct with sulfite (15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar). The crystal structure of oxidized DAAO in complex withl-lactate and of reduced enzyme complexed withd-alanine shows that the side chain of Arg285is at >5 Å from the α-CH group of the substrate. Thereby, in this position Arg285 is far way from the reactive α-hydrogen and cannot be the active site base required by a carbanion mechanism for substrate dehydrogenation. In fact, all the Arg285mutants we produced possess appreciable dehydrogenase activity (they can be anaerobically reduced by the substrate d-alanine). On the other hand, the large decrease in the rate of flavin reduction (and therefore in turnover) is quite surprising. Thek cat for R285K and R285A are decreased by about 450- and 7000-fold, respectively, in comparison to the wild-type RgDAAO (12Pollegioni L. Langkau B. Tischer W. Ghisla S. Pilone M.S. J. Biol. Chem. 1993; 268: 13850-13857Abstract Full Text PDF PubMed Google Scholar); this change is accompanied by a parallel decrease in the rate of flavin reduction (Table IV). Since k 2 is still rate-limiting, we can rule out a change in the kinetic mechanism. A main kinetic difference with respect to the wild-type DAAO is that, due to the large decrease in k 2,k 3 ≫ k 2 for the Arg285 mutants and thus the reductive half-reaction is essentially monophasic. The effect of Arg285 substitution on RgDAAO catalysis can be explained in terms of the recently proposed mechanism in which "orbital steering/interactions are the predominant or the sole important factors in catalysis."2 The perturbation of the active site in the Arg285 mutants modifies the precise substrate alignment: alteration of the reaction trajectory results in a large change in the reaction velocity (k red and k cat). A similar large effect has been also reported for the enzyme isocitrate dehydrogenase, as a consequence of small trajectory changes via substrate modification and metal co-ordination (17Mesecar A.D. Stoddard B.L. Koshland D.E. Science. 1997; 277: 202-206Crossref PubMed Scopus (200) Google Scholar). Actually, a second possibility has also to be taken into account, namely that an equilibrium form of the enzyme-substrate complex exists, in which the guanidinium side chain of Arg285 is no longer in contact with the carboxylate but it reaches a position in which it could abstract the proton from the α-carbon of the substrate. The mechanism by which substrate dehydrogenation occurs in DAAO cannot be solved solely on the basis of structural data, thus new investigations are required to rule out definitively the possibility that Arg285 could be the acid/base residue required by a carbanion-type mechanism. Our results support the concept that precise binding and orientation of the substrate is a main quantitative factor in RgDAAO catalysis. Albeit the lack of the free enzyme crystal structure, the novel hypothesis on a conformational swing of the side chain of Arg285 from the position (seen in the structure) in which it binds the carboxylate anion, to a different one (next to the flavin N(1)-C(2)=O locus) in the free enzyme form, fits with the interpretation of the whole body of data. Moreover, the model we have proposed permits a reconciliation of the apparently contradictory conclusions that arose from the three-dimensional structure and from chemical modification studies of pkDAAO (4Mattevi A. Vanoni M.A. Todone F. Rizzi M. Teplyakov A. Coda A. Bolognesi M. Curti B. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 7496-7501Crossref PubMed Scopus (275) Google Scholar, 5Mizutani H. Miyahara I. Hirotsu K. Nishina Y. Shiga K. Setoyama C. Miura R. J. Biochem. (Tokyo). 1996; 120: 14-17Crossref PubMed Scopus (125) Google Scholar, 14Nishino T. Massey V. Williams Jr., C.H. J. Biol. Chem. 1980; 255: 3610-3616Abstract Full Text PDF PubMed Google Scholar, 15Fitzpatrick P.F. Massey V. J. Biol. Chem. 1983; 258: 9700-9705Abstract Full Text PDF PubMed Google Scholar). We thank Dr. Sandro Ghisla (Universität Konstanz, Germany) for the kind hospitality of use of the stopped-flow spectrophotometer.
