Lactate Monooxygenase - The Journal of Biological Chemistry

THE

JOURNAL OF BIOLMICAI. CHEMISTRY

Vol. 269, No. 11, Issue of March 18,pp. 7982-7988, 1994 Printed in U.S.A.

Lactate Monooxygenase I. EXPRESSION OF THE MYCOBACTERIAL GENE IN ESCHERZCHZA COLI AND SITE-DIRECTED MUTAGENESIS OF LYSINE 266” (Received for publication, September 28, 1993, and in revised form, December 2, 1993)

Ute Miih$, Vincent Massey$, and Charles H. Williams, Jr.$H From the Department of Biological Chemistry, University of Michigan Medical School, Ann Arbor, Michigan 48109-0606 and the §Department of Veterans Affairs Medical Center, Ann Arbor, Michigan 48105

Lactate monooxygenase utilizes oxygen in theconversion of L-lactate to acetate,COP,and water. The gene for the enzyme from Mycobacterium smegmatis had been cloned into Escherichia coli (Giegel, D. A., Williams, C. H., Jr., and Massey, V. (1990) J. Biol. Chem. 265,66266632) and thederived amino acid sequence compared to glycolate oxidase and flavocytochrome b,, enzymes of known three-dimensional structure (Lindqvist, Y., and Brand&, C. I. (1989)J. Biol. Chem.264,3624-3628; Ma, Z. X., and Mathews, S. F. (1990) J. Mol. Biol. 212, 837-863). There is strong homology, especially around residues in the active site. The mechanism proposed for lactate monooxygenase involves an intermediate havingnegaa tive charge at theN(1)-positionof the FMN. Based onthe homology, lysine 266 is the residue suggested to neutralize that charge. Wild type enzyme and several forms of the enzyme altered at active site residues by site-directed mutagenesis have been expressed in E. coli and purification procedures developed. The properties determined forthe recombinant wild type enzyme were, in every case, the same as those previously determined for the enzyme isolated from M. smegmatis. Mutation of lysine 266 to a methionine created K266M. The semiquinone showed spectralfeaturesdifferent from those found in the wild type enzyme and was no longer thermodynamically stable. This indicates a redox potential for the enzyme-bound semiquinonelreduced flavin couple that is higher than themidpoint potential for the oxidized flavidsemiquinone couple. The two-electron redox potential was detehnined to be -180 mV at 25 “C, pH 7.0. In wild type enzyme, attack of the flavin ring by sulfite creates a negative charge at the FMN N(1bposition;. Ip K266M, the stabilization of the sulfite adduct was 17,000-fold weaker (Kd M) thaninthe wild type enzyme, with a rate of association that is lowered by 10,000-fold (ken = 1.2 M - ~s-l). The rate of reduction with L-lactate is significantly decreased in K266M. Unexpectedly, binding of substrate and inhibitors is significantly weakerin K266M than in the wild type enzyme. In all properties involving a negative charge at position N( 1)of the FMN, K266M is distinctly differentfrom wild type enzyme. This makes it quite likely that lysine 266 serves the postulated role of interacting with thisnegative charge.

-

* This work was supported by the United States Public Health Service, National Institutes of Health Grant GM 11106 (to V. M.) and National Institutes of Health Grant GM 21444 (to C. H. W.), and the HealthServicesandResearchAdministration of the Department of Veterans Affairs (to C. H. W,). The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked “aduertisernent”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1To whom correspondence and reprint requests shouldbe addressed.

L-Lactate monooxygenase is a well studied flavoprotein (EC 1.13.12.4) isolated from Mycobacteriumsmegmatis. The enzyme catalyzes the oxidation of L-lactate to acetate, carbon dioxide, and water. It is found as an octamer with identical subunits and one FMN per monomer (Sullivan et al., 1977). Although the physiological role of lactate monooxygenase in the mycobacterial species is not known (Ratledge and Stanford, 1982), the enzyme can be purified in good yields from M. smegmatis (Sullivan et al., 1977). However, expression of the gene in Escherichia coli was pursued to enable the generation of sitedirected mutants. Thereaction mechanism of lactate monooxygenase has been studied in great detail, and evidence has accumulated that describes the composition and properties of the active site (for a review, see Ghisla and Massey, 1991). For example, it i s postulated that catalysis is initiated by proton abstraction from the a-carbon atom of the substrate (Walsh et al., 1973; Ghisla and Massey, 1975, 1977); hence a base is required close to the bound substrate. Several results also indicate a positively charged amino acid residue in close proximity to the flavin N(1) position. These findings include the stabilization of an anionic semiquinone(Massey andPalmer, 19661, the tight binding of sulfite (Massey et al., 1969; Muller and Massey, 19691, and the stabilization of the benzoquinoid form of 8-mercapto-FMN (Massey et al., 1979). Developing a working model for the structure of the active site was aided by the homology of lactate monooxygenase with flavocytochrome b2 and glycolate oxidase. For both proteins the crystal structures have been solved (Lindqvist and Branden, 1989; Xia and Mathews, 1990; Xia et al., 1987) and for each of the residues implicated in binding and catalysis there is an identical residue in the amino acid sequence of lactate monooxygenase (Giegel et al., 1990). Furthermore, by flavin replacement studies with stereospecifically 3H-labeled 8-hydroxy-5deaza-FMNHz, it was shown that substrate reacts with flavin on the si-face (Manstein etal., 1986, 1988); thesamestereospecificity is apparent from the crystal structures of flavocytochrome bz and glycolate oxidase. Based on the homology with the related proteins, lysine 266 in lactate monooxygenase (Fig. 1) appears asa likely candidate for the positively charged amino acid residue that was postulated to be in the vicinity of the flavin N( 1).Expression of the gene in E . coli enabled us to test this model of the active site. Lysine 266 was replaced with a methionine, an amino acid residue that possesses similar spatial requirements but isuncharged. The mutant enzyme is referred to as K266M. This paper describes the expression of lactate monooxygenase in E. coli, the mutation of lysine 266 to a methionine, a modified purification protocol, and mechanistic studies on K266M. EXPERIMENTALPROCEDURES Materials-The E . coli strain DHSaF’ was from Bethesda Research Laboratories (BRL). The plasmid pSKM13(+)was from Stratagene (La

7982

Lactate Monooxygenase: Mutagenesis of Lysine 266

FIG.1. Representation of the active site of lactate monooxygenase. The positioning of the functional groups is derived from the crystal structuresof glycolate oxidase and flavocytochrome b, (LindqvistandBranden,1989; Xia and Mathews, 1990;Xia et al., 1987). The residues are numbered for lactate monooxygenase. NH,

Jolla, CAI. The E. coli strains BL2UDE3) and BL2l(DE3).pLysS and the plasmid pET3a were a gift from F. W. Studier (Brookhaven National Laboratory, Upton, N Y ) . Oligodeoxynucleotides were synthesized with a n Applied Biosystems model 380A DNA synthesizer by the Oligonucleotide Synthesis Center at the University of Michigan. The restriction enzymes and DNAmodifying enzymes were purchased from Boehringer Mannheim, BRL, or New England Biolabs. Yeast extract and tryptone extract were fromDifco Laboratories, Detroit, MIor BRL. Q-Sepharose was from Pharmacia LKJ3 Biotechnology Inc. L-Lactate (lithium salt), u-lactate (lithium salt), methyl viologen, benzyl viologen, imidazole, and oxalate (potassium salt) werefrom Sigma. Pyruvate (sodium salt) was purchased from Boehringer Mannheim and sulfite from J. T. Baker Chemical Co. 8-Mercapto-FMN was freshly prepared from 8-chloroFMN as described in the literature(Massey etal., 1979).DL-2-Hydroxy3-butynoic acid was a gift from Dr. S. Ghisla from the University of Konstanz, Germany. Instrumentation-Large scale cell growth was performed in a 200liter New Brunswick fermenter. Cells were harvested with a Sharples continuous flow centrifuge. A Branson sonicator, large tip, was used for lysing the cell suspension in a Branson rosette cell, 500 ml, chilled on a n icdsalt mixture. Chromatography was performed on a Waters fastprotein liquid chromatography 650E,by Millipore. Spectroscopic determination of dissociation constants and absorbance spectra were carried outwithtemperature-controlledVarianspectrophotometers,Cary model 3 and Cary model 219, or with a Hewlett Packard diode array spectrophotometer, model 8452A, all at 25 "C. Excitation and emission by spectra wererecorded with a scanning ratio spectrofluorimeter built Dr. D. P. Ballou and G. Ford at the University of Michigan. Activity Assay and Gel Electrophoresis-The activity of lactate monooxygenase was measured in a Gilson Oxygraph equipped with a Yellow SpringsInstrumentselectrode.Theassaywasperformedas described previously(Giegel et al., 1987). The purityof the enzyme was determined by SDS-polyacrylamide gel electrophoresis (Laemmli, 1970). Molecular Cloning-Both single and double-stranded DNA were sequenced using modified T 7 DNA polymerase (Sequenase) by United States Biochemicals (Cleveland, OH). a-"S-dATP (>lo00Ciimmol) was purchased from Amersham Corp. Site-directed mutagenesis followed the protocol and used the reagents from Amersham Corp. Standard cloning procedures were done according to Sambrook et al. (1989) or Ausubel et al. (1987). Computer DNA manipulations were performed with the program PC/Gene by Intelligenetics, Mountain View, CA. Expression of the Gene in E. coli and Site-directed Mutagenesis-The gene for lactate monooxygenase had been cloned into pSKM13(+) to give pDGO13 (Giegel et al., 1990) which contains a 3.1-kilobase fragment from M. smegmatis. Cutting pDG013 with XhoI and BstXI yields a 1200-base pair fragment which includes the entire lactatemonooxygenase coding region of 1182 base pairs. This fragment was blunt-ended at the BstXI site and inserted into theXhoI-EcoRV site of pSKM13(+) to yield pGM03. The plasmidpGMO3 was isolated as single-stranded DNA and served as the template for the mutagenesis reaction. In order to express the protein, the gene for lactate monooxygenase was cloned into pET3a (Studier et al., 1990).For this purpose an NdeI site (bold) was introduced at the initiation codonby mutagenesis reaction with the

7983

II

0

oligonucleotide: 5'-CTCGAGGAGACATATGAGCAATT-3'; the final construct was numbered pGM17. This r e a c t i o n s o incorporated a silent mutationof the codon for the N-terminal amino acid residue, serine (underlined), based on the transcriptional efficiency values described for P-galactosidase (Looman et al., 1987). The NdeI-BamHI fragment (1220 base pairs) from pGM17 was ligated into the appropriate site of pET3a, resulting in the expression vector pUMO1. For cloning purposes, pGM03 was grown in DH5aF'. The expression plasmid, pUMO1, was transformed into the strain BL21(DE3).pLysS. The mutagenesisof lysine 266 (codon: AAG) was performed with the oligonucleotide 5'-CGTCATCCTGATGGGCATCCAGC-3', replacingit with a methionine (underlined). Successful mutagenesis was screened by plasmid sequencing. The absence of second site mutations wasconfirmed from ClaI t o PstI. This 450-base pair fragment was ligated into the equivalent position in pGMO3 to give pGM10. An NcoI-KpnI fragment (810 base pairs) from pGMlO was ligated into pUMOl to give pUMO3. The successful construction of pUM03 was confirmed by the insertion of a MamI site (GATNN/NNATC). Cell Growth and Purification of Wild Type Lactate MonooxygenaseSix times 1 liter of Terrific Broth (Tartof and Hobbs, 1988) in 2-liter Fernbach flasks were inoculated with cells scraped from six heavily grown plates. Good induction was found only with an inoculum from plates or from cultures grown t o initial log phase. Cultureswere grown under vigorous shaking at37 "C, until theOD,,,, reached 0.8. The cells were induced with isopropyl-P-o-thiogalactopyranoside at a final concentration of 60 PMand grown for an additional P 5 h before harvesting at 7000 x g. Typically the wet weight yield was approximately 45 g of cells. The majorityofthe enzyme is found in aninsoluble form, presumably as inclusion bodies. Only the soluble fraction was isolated and purified. After freezing overnight, the cells were resuspended in 75-100 ml of buffer (25mM NaPJKE',, pH 7.0.1 mM EDTA, 1mM dithiothreitol, 0.4 mM phenylmethylsulfonyl fluoride). DNase I (10 mg) was added to reduce the viscosity. The suspension was sonicated four timesfor 1-min intervals a t maximum power. Streptomycin was added to a final concentration of 270, and the lysed cells were spun at 5,000 x g, 4 "C for 25 min. The pellet contains the insoluble enzyme and was discarded. The supernatant remained turbid and was clarified in an ultracentrifuge at 4 "C, 25,000 revolutionsimin for 20 min, followed by 45,000 revolutions/ min for 60 min. The pellet was again discarded and the supernatant dialyzed against 4 liters of 25 mM Nap,/",, pH 7.0, 0.3 mM EDTA. The dialyzed fraction was clarified by centrifugation at 25,000 x g, 4 "C for 20 min and applied to a 200-ml Q-Sepharose column. The column was washed with 3 to 4volumes of buffer (25 mM Nap,/",, pH 7.0,0.3 mM EDTA) at 10 mumin. A stepwise increase of NaC1, at 0.1 and 0.2 M was applied, followed by a gradient from 0.3 to 0.5 M NaCI, at 10 mumin over 100 min. Lactate monooxygenase eluted around 0.35 M NaCI. All enzyme-containingfractions were pooled and precipitated with 85% ammonium sulfate. Theammoniumsulfatepelletwasresuspendedinaminimum amount of 1M NaOAc, pH 5.4, anddialyzed against the same. Centrifugation after this first crystallization step gave a white pellet under a

7984


translucent yellow layer. m e r resuspension in 10 nm ImCI,' pH 7.0, only the yellow layer went into solution. The suspension was spun again and the white pellet discarded. The supernatant was crystallized a second time by dialysis against 1 M NaOAc, pH 5.4. The crystalline dialysate was then stored at"C4 in the dark. For most experiments an at 14,000 aliquot of thecrystallinesuspensionwascentrifuged revolutiondmin for 10 min, the pelletdissolved in 10 m ImC1, pH 7.0, and desalted over a G-25 column equilibrated with the same buffer. Experiments were performed in 10 mM ImCl, pH 7.0, a t 25 "C unless noted otherwise. Cell Growth a n d Purification of K266M Lactate MonooxygenaseThe K266M mutant is expressed well, as seen by running whole cells of an induced culture on a n SDS-polyacrylamide gel, stained with Coomassie Blue.However, the amountof enzyme found in thesoluble fraction 300 400 500 600 700 is substantially less than for wild type (all other mutations investigated so far haveyielded amounts of soluble enzymein the range of that seen Wavelength (nm) with wild type enzyme). Consequently K266M was grownin a 200-liter fermenter. Six times1liter of medium were inoculated from a plate and FIG. 2.Spectral properties of JS266M.Solid line,oxidized enzyme; grown for about 3 h until the OD,,, nm reached 0.3. These 6 liters of dotted line, substrate-reduced K266M. The enzyme concentration was initial log-phase culture were then used to inoculate the 200 liters of 1.1 x M as determined by flavinextinction(10 m ImCI, pH 7.0). medium. Thecells were aeratedvigorously at 37 "C and induced with 20 p~ isopropyl-P-o-thiogalactopyranosidewhen their OD,,, reached 0.8. from the side-armof the anaerobic cuvettet o a The cells were grown for an additional 4 h and harvested. The cell paste the inhibitor was added final concentration of 1.5 m. The changes of absorbance and fluoreswas frozen at -20 "C before breaking. cence properties were measured with time. Due to the low enrichment for K266M in the soluble fraction, the Redox Potential-The redox potential was determined by anaerobic purification protocol was modified. The changes make useof the therof a mal stability of this mutant. Cells at a wet weight of 230 g were resus- additions of pyruvate to substrate-reduced enzyme in the presence known amount of lactate. The ratioof reduced to oxidized enzyme was pended in 400 ml of lysis buffer (10nm ImCI, pH 7.0,0.2 M KCl, 10 m MgCl2, 1 n m dithiothreitol, 1m phenylmethylsulfonyl fluoride) and 20 calculated by the absorbance at 450 nm, using the determined extincmg of DNase I was added. The suspension was sonicated on ice at least tion coefficient. The ratio of lactate to pyruvate was calculated by the 5 times for 1-min intervals at maximumpower. The lysate was centri- known amounts added. Preparation of 8-Mercapto-FMN Lactate Monooxygenasefuged at 25,000 x g, 4 "C for 1h toremove both inclusionbodies and cell membrane. The supernatant was then incubated for 60 min a t 37 "C Apoprotein of K266M was preparedby acid ammonium sulfate precipiwith gentle shaking. A substantial amountof protein precipitated and tation as describedfor wild type lactatemonooxygenase (Choong et al., 1975). The procedure was modified in order to improve the yield of was removed by a second centrifugation. The supernatant wasconcenreconstitutable protein, since only small quantities of the mutant entrated by precipitation of protein at 90% saturation with ammonium zyme were available. K266M was subjected to two acid ammonium sulfate, followed by several crystallization steps against 1M NaOAc, pH 5.4, as described previously. This procedure requiredclose monitoring of sulfate precipitations, followed by one neutral ammonium sulfate preall fractions, sometimes calling for adjustments to salvage enzyme thatcipitation. The pellet was then dissolved in a 2-fold excess of 8-merwas not found in the expected fraction. Monitoring was done by gel capto-FMN adjusted to pH 8.0. For wild type lactate monooxygenase 50 TrisC1 at a pH of electrophoresis since the low catalytic activity of the mutant enzyme reconstitution of the apoenzyme was optimal in mM 8.0 (datanot shown). precludes the use of activity measurements. Extinction Coeficient-The extinction coefficients for the mutant forms of lactate monooxygenase were determined by denaturing the RESULTS protein with SDS. A 10-pl aliquot of 10% SDS was added t o 1 ml of Expression-The gene for lactate monooxygenase did not exenzyme solution in 10 IILM ImC1, pH 7.0, at 25 "C. The extinction coefficient for the flavin bound to lactate monooxygenase was determined press well under control of the tac promoter of the plasmid based on the amount of released FMN. The concentration of free FMN pKK223.3, nor was there a satisfactory amount of activity uswas calculated usingan extinction coefficient of 12 500M" cm" a t 445 ing the T7 promoter in Bluescript pSKM13(+). Performing a nm. silent mutation on the N-terminal amino acid to increase the Reductive Halfreaction-The enzyme solution was treated with altranscriptional efficiency value (Looman et al., 1987) also did ternate cycles of evacuation and oxygen-free argon in an anaerobic not improve the amount of expressed enzyme. It appears that cuvette. Once the solution was made anaerobic,L-lactate was tipped in from the side-arm and the changes at 450 nm with time were measured additional translation signals arerequired. The nucleotide sequence 5' to the initiation codon was mutated by polymerase in the spectrophotometer. Photoreduction-Lactate monooxygenase was photoreduced accord- chain reaction to correlate with the ribosome-binding site deing to the method described in the literature (Massey and Hemmerich,scribed for the T7 gene 10 product (Olins et al.,1988; Olins and 1977; Massey and Hemmerich, 1978). However, glycine was used inRangwala, 1989). This increased theamount of expression stead of EDTA, since EDTA photochemically breaks down to glyoxalate (Muh, 1993). So the gene was further cloned into the PET 3a which is a substrate for lactate monooxygenase. The reaction mixture containedenzymein10 m ImCI, pH 7.0,0.1 M glycine, and 1 1.1~ vector (Studier et al., 1990) which incorporates all the tran5-deazaflavin. In the dark the mixture was treated with alternatecycles scriptional and translational signalsof the T7 system. At least 5% of the total cell protein is the expressed target gene; howof evacuation and flushing with oxygen-free argon. Light irradiation was performed with visible light from a sun gun (Smith Victor Corp., ever, over 95% of this is found in an insoluble form. Attempts Griffith, IN) at an intensityof about 6 x 10, erg s-l cm-2. were made to avoid the formation of inclusion bodies or to Stability of the Semiquinone-The thermodynamic stability of the purify active protein out of the insoluble fraction. Neither of semiquinone species was tested by photoreducing the mutantenzyme to full semiquinone. Thenbenzyl viologen was added toa final concentra- these approaches proved successful. Consequently, the purification of the soluble protein fraction was pursued. tion of 6 x M to act as a mediator between the flavin semiquinone Purification-Typically 80 mg of wild type enzyme was puin the octameric protein. The absorbance changes werefollowed for 15 h at 25 "C. rified from 6 liters of induced cultures. In all the properties Mechanism-based Inactiuation-To test the reactivity of the lactate examined, the recombinant enzyme behaved just as the enmonooxygenase mutant with 2-hydroxy-3-butynoic acid (Walsh et al., zyme purified from M. smegmatis. The 200-litergrowth of 1972; Ghisla et al., 1976). the enzyme solution was made anaerobic, and

K266M gave 1.2 kg of wet weight cell paste. Approximately 2.5 mg of protein was purified from 100 g of this paste. The abbreviations used are: ImCl, imidazole-HC1 buffer E,,, oxiSpectral Properties-The spectral properties of K266M are dized enzyme; Esq,enzyme semiquinone;Ere+two-electron reduced enshown in Fig. 2. The oxidized enzyme has a maximum at 465 zyme.


7985

0.15 378

0

300

400

500

600

300 500

700

nm (458 nm in wild type enzyme) with well developed shoulders, which is separated by a deep trough from an almost resolved peak at 360-372 nm. The ratio A28dA465is 10.3, and the extinction coefficient at 465 nm was determined to be 11,500 M - ~cm-'. In contrast to the non-fluorescent wild type enzyme, the oxidized species of K266M is weakly fluorescent with excitation maximaat 465 and 370 nm and anemission peak at 530 nm. These properties, and other physicochemical characteristics detailed in the following sections, are only marginally different from those of wild type enzyme and indicate strongly that the maineffects of the mutations areon the kinetic properties of the enzyme, rather than on the protein structure. Reduced K266M has a single broad maximumat -400 nm and shows negligible fluorescence, again different from wild type enzyme which is stronglyfluorescent inthe reduced form (Ghisla et al., 1974). Given that a positive charge adjacent to the flavin has been removed by the mutation, it is not surprising that minor changes in the fluorescence properties of the flavin are exhibited. Themutant enzyme forms an anionic semiquinone (Fig. 3). However, the spectral features are distinctly different from wild type enzyme. There is a very broad peak at 495 nm with a shoulder at -530 nm and a maximum at 378 nm. The sharp peak at -400 nm in wild type enzyme has been replaced by a small shoulder. Stability of the Semiquinone-The thermodynamic stability of the K266M semiquinone was tested by adding benzyl viologen after photoirradiation. Fig. 3 shows the result of this experiment. Unlike the wild type enzyme, the semiquinone is not thermodynamically stable, rather it disproportionates to an equimolar mixture of oxidized and reduced enzyme. Within 15 h all absorbance a t 530 nm had disappeared and the 465 nm peak had increased, indicating a mixture of oxidized and reduced species. Upon opening the cuvetteto air, full absorbance was recovered. Redox Potential-The redox potential for two-electron reduction, E o d r e d , was measured by anaerobic titration of the enzyme with pyruvate after reduction by L-lactate (Fig. 4). The redox potential was determined tobe -180 mV at pH 7.0, 25 " C , from the Nernst plot (Fig. 4, inset). The calculated overall redox potential for two-electron reduction in the wild type enzyme is -149 mV underthesame conditions(Stankovich and Fox, 1983). The more negative potential of the mutant enzyme is consistent with the removal of a positive charge in theprotein interacting preferentially with the reduced flavin (Clark, 1960). Reductive HaZf-reaction"K266M is reduced by L-lactate very slowly. A direct plot of the observed rate against lactate con-

600

700

Wavelength (nm)

Wavelength (nm) FIG.3. Thermodynamic instability of the semiquinone.The oxidized enzyme (solid line)was reduced by photoirradiation t o maximal semiquinone (dotted line).On addition of benzyl viologen, the semiquinone disproportionated to a mixture of oxidized and reduced enzyme. The spectra were taken 16 min,9 h, and 15 h after the addition of the electron mediator (dashed lines).The experiment was performed in 10 mM ImCl a t pH 7.0, 25 "C.

400

FIG.4. Redox titration against the L-lactate-pyruvatecouple. The oxidized enzyme (solid line) was reacted with 50 mM L-lactate to generate the fully reduced species (dotted line). Pyruvate was added under anaerobic conditions, and the following concentrations are shown (dashed lines):11, 32, 52, 77, and 121 mM. A Nernst plot of the calculated data is depicted in the inset. The slope is 1.0which is equal to the theoretical value; and the redox potential for E,,,,, was measured as -180 mV in 10mM ImCl at pH 7.0,25 "C. The enzyme concentration in this experiment was 1.1 x lo-" M as determined by flavin extinction.

0.002

--

b W

Ul

v

::

9

0.001

0

0.1

0.2

[Lactate] (M) FIG.5. Reductive half-reaction with L-lactate. K266M was reacted anaerobically with L-lactate in 10 mM ImCl a t pH 7.0, 25 "C. The observed rate was plotted directly against substrate concentration and shows either a linear relationship witha slope of 10" M" s-l, or saturation behavior with a limiting rate of ssl and a Kd for L-lactate of about 1 M.

centration can be interpreted as a straight line which passes through the origin (Fig. 5) and with a slope of s-l. Alternatively, the relationship can be interpreted as showing saturation behavior with a Kd -1 M and a limiting rate of s-'. Under the same conditions wild type enzyme shows saturation behavior in the rate of reduction by L-lactate, with a limiting rate of 230 s-l and an estimated Kd of 5 x M. The binding characteristics and kinetic properties of K266M are compared with those of wild type enzyme in the third paperof this series. Binding of D - L U C t U t e and Oxalate"K266M binds D-lactate weakly (Fig. 61, with a Kd of approximately 0.3 M (wild type M). The interactionwithoxalate is similarly enzyme: 1.7 x weak, and the spectral changes are small, so that the Kd can only be estimated to be approximately 0.1 f 0.04 M. Oxalate does, however, bind in a fashion that permits theformation of a photoadduct, as with wild type enzyme. In the presence of absorbance at 465 0.17 M oxalate and upon light irradiation, the nm decreases. In contrast to wild type enzyme, the covalent adduct (presumably the N(5bcarbonyl flavin) is non-fluorescent. At 25 "C the photoadduct decays at a rate of 3.3 x lo-" s-l, about two orders of magnitude faster than the wild type ens-l; Ghisla and Massey, 1975), returnzyme adduct (5.1 x ing the enzyme to its initial oxidized flavin state.


7986

300

I 0

a

1 /[Sulfite] ("1)

J

20

40

Ih

FIG.6. Binding of o-lactate. The observed changesat 520 nm and concentrations of D-lactate are shown in a double-reciprocal plot to give a Kd of approximately 0.3 M. The experiment was performed in10 m ImCl at pH 7.0, 25 "C.

r ' / f1

'

0.006/-

1 / [D-lactate] ("1)

.2M" s'

9

,

0.002 intercept:

Binding of Sulfite"K266M binds sulfite with almost complete abolition of the visible absorption spectrum, presumably due to formation of the flavin N(5)-sulfite adduct, as with wild type enzyme (Masseyet al., 1969; Mullerand Massey, 1969). A Kd of 9 x M was determined by equilibrium titration, Fig. 7a (wild type enzyme: Kd = 5.6 x M). The rate of association was slow enough to be followedafter theindividual additions of sulfite and a plot of Kobs uersus the sulfite concentration could be determined, Fig. 7b. From they intercept,koflis determined as 1.4 x s-l (wild type enzyme: 7.2 x s-l), and the slope gives k,, with a value of 1.2 M - ~s-l (wild type enzyme: 1.3 x lo4 M - ~ 6-l). Thus, the calculated K d is 1.2 x M, in reasonable agreement with the value obtained from the equilibrium titration. Reaction with 2-Hydroxy-3-butynoic Acid-The mutant enzyme reacted slowly with 2-hydroxy-3-butynoic acid, at a rate which parallels the decreased reactivity with substrate. In the presence of 1.5 mM inhibitor, the reaction was complete within 1h (Fig. 8). The fluorescence emission is at 490 nm, compared with wild type enzyme, which shows an emission maximum at 502 nm (Ghisla et al., 1976). However, in contrast to the wild type adduct,which decays with a half-time of several hours, the adduct with the mutantenzyme is very stable. It is essentially unchanged after 7 days at room temperature and exposure to air. 8-Mercapto-FMN-The spectra of wild typeenzyme and K266M after addition of 8-mercapto-FMN to the apoenzymes are shown after 20 min on ice (Fig. 9a) and after an additional 30 min a t room temperature (Fig. 9b).An initial reaction takes place which is similar for both wild type enzyme and K266M. Both show a peak a t 525 nm indicative of free 8-mercapto-FMN or the bound thiolate species, as well as a substantial shoulder around 600 nm, typical for the benzoquinoid form (Fig. 9a). When the reconstitution mixtures were then warmed to room temperature, the wild type sample changed to its stable blue form, indicative of the bound benzoquinoid species. The K266M sample, however, released all the initially bound flavin as indicated by the spectral properties of unbound 8-mercapto-FMN (Fig. 9b). After ultrafiltration, the majority of the flavin was found in the flow-through of the K266M sample, while the flow-through of wild type enzyme showed negligible absorbance (results not shown).

k,!, = 1.4 x

0

1 0 3 s-'

0.002

0.004

[Sulfite] (M) FIG.7. Reaction with sulfite. The results of the equilibrium titration of K266M with sulfite in 10 m ImCl at pH 7.0,25"C is depicted in a . Upon reacting with sulfite the absorbance a t 465 nm decreased; from the double-reciprocal plot the Kd was determined to be about 9 x M. b, the rate of reaction was dependent on sulfite concentration and a direct plot is drawn. From the slope, the rate of association can be determined as 1.2 M - ~s-l. The y intercept gives the off-rate to be1.4x 10-3 S-1. 0.12

300

,

,

400

I

,

500

,

I

600

I

,

700

Wavelength (nrn) FIG.8.Reaction with 2-hydroxy-3-butynoicacid. The initialoxidized enzyme is shown, as well as successive time points at 3.3, 5.5,8, 15, 25, and 66 min after addition of 1.5 mM inhibitor in 10 II~MImCl at pH 7.0,25"C. The enzyme concentration was 1.0 x M a s determined by flavin extinction.

positive charge are found in the stabilization of an anionic semiquinone, the stabilization of the benzoquinoid form of 8-mercapto-FMN, and the tightbinding of sulfite to the flavin N(5)-position. The charged amino acid residue presumably also plays a role in stabilizing the reaction intermediate, the reduced flavin anion.It was expected that theseproperties would be affected upon replacing lysine266 with an uncharged amino acid residue. Thespectral properties of K266M indicateasignificant change in theflavin environment.The interactions that quench the fluorescence in the oxidized wild type enzyme have been DISCUSSION affected such that fluorescence is emitted in the mutant enThe mutantK266M was designedto test the postulate that a zyme. In contrast, the strong fluorescence found in the wild positively charged amino acid residue in the vicinity of the type enzyme for the photoadduct of the oxalate-enzyme comflavin N(1) isresponsible for neutralizing a negative charge at plex is not present in K266M. Additionally, while the reduced that position, as described below. The effects of this protein flavin of wild type enzyme is strongly fluorescent, the corre-


7987

carbanion formation. This species would be the flavin N(5)substrateadduct with a negativelycharged N( 1)-position, which in turnmay be stabilized by interaction with lysine 266. 0.6 1: I Removal of this salt bridge could result in a substantially decreased rate of reduction. However, catalysis is probably affected also by the diminished binding capacity of K266M. There is eitherno indi'cation of a binding step for L-lactate in the reductive half-reaction or a Kd for L-lactate of -1 M, an observation which cannot be explained readily. Weak binding is also found for o-lactate, an inhibitor of the wild type enzyme, and for oxalate, a transition state analog. However, oxalate binding still reflects the decrease in the rate of reduction, a correlationthat was found for other investigated mutations (Muh et al., 1993b). The effect of the mutationon binding of sulfite met the initial expectations. K266M binds sulfite 17,000-fold weaker than wild type enzyme, underlining theeffect of the lysine residue in :. .. 0 1 stabilizing a negative charge at the flavin N(1). This impor300 500 700 tance is reflected mainly in the rateof adduct formation, which Wavelength (nm) is 10,000-foldlower than thatof the wild type enzyme, implying FIG.9.Reconstitutionof apoK266M with 8-mercapto-FMN.The that the positive charge of lysine 266 is more important in addition of 8-mercapto-FMN to the apoenzymes of wild type enzyme stabilization of the transition state, rather than of the N(5)and K266M are shown, both after 20 min on ice (a)and after an addi- sulfite complex itself. tional 30 min at room temperature (wt, -; K266M, - - - - -1 ( b ) . Conclusions can only be drawn tentativelyas to the extent to which the lysine residue participates in the stabilization of a benzoquinoid 8-mercapto-FMN. Reconstitution of the apoensponding mutant enzyme form is essentially non-fluorescent. The mutant enzyme forms a n anionic semiquinone on one- zyme is a two-step process with an initial fast binding, followed electron reduction. However, the spectral features are signifi- by a slow conformational change, in which the protein assumes cantly different from those of the wild type enzyme, as well as the proper folding around the flavin (Choong et al., 1975). It from other mutationsexamined in thesucceeding papers (Muh appears as if the apoenzyme of K266M is able to bind 8-meret al., 1994a, 1994b). More importantly, the semiquinone is not capto-FMN initially, but cannot carry out the necessary conforthermodynamically stabilized, as it is with the wild type en- mational change. Rather, it denatures andreleases any bound zyme, butis only kinetically favored upon photoreduction. flavin. It isconceivable that theamino acid residues around the Thus, the postulate that a protein positive charge in thevicin- flavin are rearranged such that the binding site can no longer ity of the flavin N( 1)-position is involved in thethermodynamic accommodate the mercapto derivative. It does, however, seem stabilization of the flavin anionic semiquinone (Massey et al., as if the initial bindingfavors the benzoquinoid species as 1979) is borne out by the data. From the thermodynamic deshown in the similar spectra for wild type enzyme and mutant stabilization of the semiquinone, it can be concluded that the enzyme. redox potential for the semiquinone to fully reduced species, The assumption that themechanism by which K266M cataEsqlred, is higherthanthe redox potential for oxidized to lyzes the reaction with L-lactate has not been substantially semiquinone, Eodsq.This is in sharpdistinction to the situation changed is supported by its ability to react with 2-hydroxy-3with wild type enzyme, where Esqlred has been estimated as butynoic acid. The rate of reaction parallels the rate atwhich -231 mV compared with -67 mV for Eoxlsq(Stankovich and the mutant enzyme is reduced by the L-lactate. Again an effect Fox, 1983). is seen on the spectral properties of the flavin. The covalent Therate of reductionwith L-lactate issubstantially de- adduct shows a fluorescence emission maximum at 490 nm, creased in K266M compared with that with the wild type en- compared with 502 nm found in the wild type enzyme. A rezyme. The potential for two-electron reduction was measured markable difference is in the stability of the adduct, which as -180 mV at pH 7.0,25 "C. This isonly 30 mV lower than that decays with a half-time of hours in wild type enzyme, but is for wild type enzyme (Stankovich and Fox, 1983). The redox stable for many days with K266M. This suggests that theposipotential for the lactate/pyruvate couple is -189 mV (Clark, tive charge of Lys-266 in wild type enzyme contributes insome 1960). Consequently, the decrease in catalytic rate cannot be way to the instabilityof the N(5)-C4a adduct formed from 2-hydue to the altered redox potential of the flavin, but must be an droxy-3-butynoic acid (Schonbrunn et al., 1976). effect on the ability of the enzyme to bind substrate or to staIn flavocytochrome 6, an homologous mutation, K349R, had bilize reaction intermediates. The very slow reduction of the led to completely inactivated enzyme (Reid et al., 1988). It was enzyme flavin by even high concentrations of lactate, coupled argued thatpacking of the amino acid residues near the flavin with the poor binding of substrate, makes the mutantenzyme pyrimidine was tight and may not be able to accommodate a virtually inactive, and precludes the meaningful determination bulky replacement of the lysine residue without loss of activity. of Kcat and K, values. The K266M mutation underscores the homology between flaIt is generally agreed that catalysis by lactate monooxygen- vocytochrome bz and lactatemonooxygenase in the substantial ase is initiated by proton abstraction from the a-carbon of the effect on the ability to be reduced by substrate. The fact that substrate. ThepK, of the a-carbon of lactate free in solution is residual activity is measuredin lactate monooxygenase K266M presumably in the order of 20. The active site base is postulated may reflect the effect of a substitution with the smallermethito be a histidine, for which the pK, of the free amino acid is -6 onine rather than an arginine. Itmay also reflect the distinct (Ghisla and Massey, 1975, 1977). Various interactions can be differences to be expected for the active sites of the two eninvoked to arguea loweringof the pK, of the a-carbon hydrogen zymes. K266R was generated for lactate monooxygenase and upon binding at the active site. One such interactioncould be was expressed well; however, an even lower yield was found in the stabilization of the intermediateimmediately following the the soluble fraction. It should be interesting to attempt a better I.

7988


expression of soluble K266R or a less bulky substitution suchas Giegel, D.A., Massey, V., and Williams, C . H., Jr. (1987)J . Bid. Chem. 262, 5705-5710 K266A. Giegel, D. A., Williams, C. H. Jr., and Massey, V (1990)J . Biol.Chem. 265, For the active site mutations that aredescribed in the suc66264632 ceeding papers, a linear relationship was determinedbetween Laemmli, U. IS. (1970)Nature 227,680-685 Lindqvist, Y.,and Branden, C . I. (1989)J. B i d . Chem. 264,3624-3628 the rate of reduction and the stabilization of the transition Looman, A. C., Bodlaender, J., Comstock, L. J., Eaton, D., Jhurani, P., deBoer, H. state analog, oxalate (Muh et al., 1993b). H290Q is the only A,, and van Knippenberg, P. H. (1987)EMBO J . 6, 2489-2492 mutant enzyme that does not fit this relationship, due to the Manstein, D. J., Pai,E. E , Schopfer, L. M., and Massey, V. (1986)Biochemistry 27, 68074816 removal of the active site base. In K266M, the dissociation Manstein, D. J., Massey, V., Ghisla, S., and Pai, E. F. (1988)Biochemistry 27, constant for oxalate is determined to be approximately 0.1 M. 2300-2305 Despite the large margin of error, this value still correlates Massey, V., and Hemmerich, P. (1977)J. Biol. Chem. 252, 56125614 Massey, V., and Hemmerich, P. (1978)Biochemistry 17, 9-17 reasonablywith the established linear relationship for the Massey, V., and Palmer, G . (1966)Biochemistry 5,31813189 other mutant enzymes. This supports the assumption that the Massey, V., Muller, F., Feldberg, R., Schuman, M., Sullivan, P. A., Howell, L. G., Mayhew, S. G., Matthews, R. G., and Foust, G . P. (1969)J . B i d . Chem. 244, mutation has not been disruptive in a significant fashion and 39994006 the effects measured can be indeed attributed to the lysine to Massey, V., Ghisla, S., and Moore, E. G. (1979)J. Biol. Chem. 254, 964&9650 Miih, U. (1993)Lactate oxidase: Expression of the Gene in Escherichia coli and methionine substitution. Studies of the Reaction Mechanism throughActcue Site Mutations.Ph. D. thesis, In all properties involving a negative charge at the flavin University of Michigan N(1)-position, K266M is distinctly different from wild type en- Miih, U., Williams, C. H., Jr., and Massey, V. (1994ajJ. Biol. Chem.269,7989-7993 zyme. Thus, the data support the postulate thatlysine 266 is Miih, U., Williams, C. H., Jr., and Massey, V. (1994b)J. Biol. Chem.269,7994-8000 F., and Massey, V. (1969)J. Biol. Chem. 244,40074016 the amino acid residue which is responsible for stabilization of Muller, Olins, P. O., and Rangwala, S. (1989)J. Biol. Chem. 264, 16973-16976 anionic forms of the flavin bearing a negative charge at the Olins, P. O., Devine, C . S., Rangwala, S. H., and Kavka,K. S. (1988)Gene (Amst. 73,227-235 N(1)-position. Acknowledgments-We thank Dr. D. A. Geigel for help in the early stages of this work. REFERENCES Ausubel, F., Brent, R., Kingston,R. E., Moore, D. D., Seidman, J. G., Smith, J. A,, and Strubel, K. (1987)Current Protocols in Molecular Biology, Greene Publishing Associates andWiley Interscience, New York Choong, Y.S., Shepherd, M. G . , and Sullivan, P. A. (1975)Biochem. J. 145, 3745 Clark, W. M. (1960)Oxidation Reduction Potentialsof Organic Systems, Williams & Wilkins, Baltimore Ghisla, S., and Massey, V. (1975)J. Biol. Chem. 250, 577-584 Ghisla, S., and Massey, V. (1976)in Flavins and Flavoproteins(Singer, T.P., ed) pp. 213-217, Elsevier Scientific PublishingCompany, Amsterdam Ghisla, S., and Massey, V. (1977)J. B i d . Chem. 252,67294735 Ghisla, S., and Massey, V. (1991)in Chemistry and Biochemistry ofFlauoenzymes (Muller, F., ed) Vol. 11, pp. 243-289, CRC Press, Boca Raton, FL Ghisla, S., Massey, V., Lhoste, J. M., and Mayhew, S. G . (1974)Biochemistry 13, 589-597 Ghisla, S., Ogata, H., Massey, V., Schonbrunn, A,, Abeles, R. H., and Walsh, C. T. (1976)Biochemistry 15, 1791-1797

Ratledge, C., and Stanford,J. (1982)The Biology OfMycobacteria,Academic Press, London Reid, G. A,, White, S., Black, M. T., Lederer, F., Mathews, F. S., and Chapman, S. K. (1988)Eur J. Biochem. 178, 329-333 Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989)Molecular Cloning: a Laboratory Manual, Cold Spring Harbor Laboratory Press,Cold Spring Harbor, NY Schonbrunn, A., Abeles, R. H., Walsh, C. T., Ghisla, S., Ogata, H., andMassey, V. (1976)Biochemistry 15, 1798-1807 Stankovich, M., and Fox, B. (1983)Biochemistry 22,44664472 Studier, W. F., Rosenberg, A. H., Dunn,J . J., and DubendorfT, J . W.(1990)Methods Enzymol. 185,60-89 Sullivan, P. A,, Choong, Y. S., Schreurs, W. J., Cutfield, J. F., and Shepherd, M. G . (1977)Biochem. J. 165, 375-383 Tartof, K. D., and Hobbs, C. A. (1988)Focus 9, 12 Walsh, C., Schonbrunn, A., Lockridge, O., Massey, V., and Abeles, R. H. (1972)J. B i d . Chem. 247,60046006 Walsh, C., Lockridge, O., Massey, V , and Abeles, R. (1973)J . B i d . Chem. 248, 7049-7054 Xia, 2. X., and Mathews, S. F. (1990)J. Mol. B i d . 212, 837-863 Xia, Z. X., Shamala, N., Bethge, P. H., Lim, L. W., Bellamy, H. D., Xuong, N. H., Lederer, F., and Mathews, S. F. (1987)Proc. Natl. Acad. Sci. U. S. A. 84,26292633