Eur. J. Biochem. 82, 143-154(1978)
A Spectrophotometric Rapid Kinetic Study of Reactions Catalysed by Coenzyme-BIZ-Dependent Ethanolamine Ammonia-Lyase Michael R. HOLLAWAY, Hugh A. WlIITE, Keith N. JOBLIN, Alan W. JOHNSON. Michael F. LAPI'ERT, and 0. Caryl WALLIS Department of Biochemistry, University College London and School of Molecular Sciences, University of Sussex (Received January 11, 1977)
1. Spectral changes during the transient states and the steady states of the reactions catalysed by the adenosylcobalamin-dependent ethanolamine ammonia-lyase with L-2-aminopropanol and 2-aniinoethanol as substrates were followed by using a rapid-wavelength-scanning, stopped-flow spectrophotometer (800 spectra per second and each spectrum 340- 560 nm). 2. During the steady state of the reaction with the 'poor' substrate, L-2-aminopropanol (k,,, = 1 s-l), under conditions of V , the spectrum of the coenzyme closely resembled that of cob(I1)alaniin. With the 'good' substrate, 2-aniinoethanol (k,,, z 140 s - I ) , under the same conditions, the spectrum corresponded to a mixture of 58 0,; cob(I1)alamin and 42 o/, cob(I1I)alamin. The latter may be either unchanged adenosylcobalamin or a coenzyme derivative with a covalent bond between a substrate carbon and the cobalt atom. Thc magnitude of a reported kinetic deuterium isotope effect favours the second possibility. There was no evidence for the presence of a cob(1)alamin intermediate. 3. The transient phases of the reactions were dependent on the order of mixing of reactants. When enzyme was placed in one syringe of the stopped-flow apparatus and coenzyme and substrate in the other, the cob(1I)alamin intermediate formed slowly (first-order rate constants of about 5 s-' with L-2-aminopropanol as substrate and about 2 s-l with 2-aminoethanol). When the enzyme and coenzyme were placed in one syringe and the substrate in the other, the formation of the cob(1I)alamin occurred much more rapidly (90 s - ' with L-2-aminopropanol and greater than 300 s-' with 2-aminoethanol). It is concluded that binding of adenosylcobalamin to the enzyme is followed by a slow change in the conformation of the enzyme molecule to give a catalytically active species (there is no observable change in the spectrum of the coenzyme concomniitant with this process). The formation of active holoenzyme is slow on the time scale of subsequent catalytic steps when 2-aminoethanol is the substrate. 4. The spectra of the intermediates formed in the reactions with both substrates, the rates of their formation and their lifetimes allowed a detailed kinetic analysis of the reactions in terms of a threestep mechanism involving binding of substrate. cob(I1)alamin formation ( k + z step) and cob(I1)alamin breakdown ( k + 3step). I n catalysis the difference between the substrates is expressed mainly in the /i+, step (240 s-l and about 1 s-l respectively for 2-aminoethanol and L-2-aniinopropanol, the respective k + values being 336 s- and 90 s '). ~
We have continued our studies [ l ] on the mechanism of action of the adenosylcobalatnin-dependent ethanolamine ammonia-lyase from Clostridiian sp., as approached by the measurement of the spectra of reaction intermediates by stopped-flow, rapid-scanning spectrophotometry [2]. These experiments in-
volve mixing enzyme and substrate solutions in about 3 ins and then recording spectra over a 200-nm range at the rate of 800 s-'. Ethanolamine ammonia-lyase catalyses the formation of acetaldehyde and ammonia from 2-aminoethanol or of ammonia and propionaldehyde from L-2-aminopropanol, in reactions which each involve the of an amino group (which may be proThis paper is Part V of the series C o r n ~ ~ , ) ~ c , - B , ~ - ~ c , / ~ r , ~ c c / rmigration /rt tonated. Scheme 1 ) between substrate carbon atoms. Rf,ocriorls (for Part 1V see [9]). The enzyme molecule. molecular weight 520000, A h h ~ ~ i ~ t i AdoCbl. ~ i i s . adenosylcobalamin; ESR, electron spin contains two active sites which appear to function resonance. EKJWK,. Ethanolamine ammonia-lyast: (EC 4.3.1.7). independently [3,4].
144
Ethanolamine Ammonia-lyase: a Transient Kinetic Study HO R-CH,
I
+ I
NH,
CH,-CH,
+ I
HO
NH,
I
R-CH,
CH,-CH,
AdoCbl
1
PH
HO
$H-CH,
I
I
I
t
I I
I
/"" CH,-CH R-CH,
AH,
I N
R-CH,
HO R-CH,
+
NH,
r-CH2 I
I
Scheme 1. Mechanistic .scher~ir,/or. the r'ructioti ccitalysrtl hy eihanolanzinr urnnioriici-/)use with 2-aniinoeihanol as suhsirate R - CH2 - = 5'-adenosykobdkimin The catalytic steps of the reaction are believed to proceed by a mechanism which may (stippled arrows) or may not involve substrate-metal binding during rearrangement. A similar scheme can. in principle, be drawn for L-2-aminopropanol as substrate
In a previous communication [l], it was demonstrated that, with L-2-aminopropanol as substrate and AdoCbl [containing Co(III)] as cofactor, the characteristic spectrum of a Co(I1) form of the coenzyme [designated as cob(II)alamin, formerly as B I ~ , ]appeared during the enzymic reaction in a process with a first-order rate constant of about 3 s-l. As the catalytic centre activity for ethanolamine ammonia-lyase with L-2-aminopropanol as substrate is between 1 and 2 s-l [ 5 ] , the rate of formation of the cob(I1)alamin species was consistent with it being an intermediate on the catalytic pathway. This observation was in agreement with previous freeze-quenching ESR spectroscopic studies with the same substrate and cofactor [6], which indicated that a paramagnetic Co(I1) species formed at a kinetically competent rate after mixing enzyme with AdoCbl and L-2-aminopropanol. Given that the Co(I1) species of the coenzyme forms at a kinetically competent rate during catalysis, then a radical rearrangement mechanism seems likely for the reaction catalysed by ethanolamine ammonia-lyase (see Scheme l ) , at least with L-2aminopropanol as substrate, although the intervention of a carbonium ion rearrangement cannot be finally excluded on the evidence available at present. The catalytic centre activity for ethanolamine ammonia-lyase with 2-aminoethanol as substrate is about 140 s - ' [4] and this is 100-fold greater than that with L-2-aminopropanol, so care must be exercised in extrapolating conclusions reached by consideration of results obtained by using the 'poor' substrate to the reaction with the 'normal' substrate. Nevertheless,
it has been shown [l,71, that a cob(I1)alamin species of the coenzyme is present during the reaction catalysed by ethanolamine ammonia-lyase with 2-aminoethanol as substrate. However, there has been no demonstration of the kinetic competence of the cob(I1)alamin species as an intermediate in this reaction and part of the present study was directed towards the investigation of this point. In addition, results are presented which demonstrate that ethanolamine ammonia-lyase undergoes a structural change on binding coenzyme, to give a catalytically active species, and that, with both L-2-aminopropanol and 2-aminoethanol as substrate, this change is much slower than subsequent turnovers. MATERIALS AND METHODS Ethanolamine ammonia-lyase was purified from cultures of Clostridium sp. by the method of Kaplan and Stadtman [8] with the modifications described by Joblin et ul. [9]. Bound cobalamin derivatives were removed from the enzyme by the procedures described in these papers. Molar concentrations of enzyme active sites in reactant solutions were calculated from measurements of enzyme activity, assuming that the specific activity of the enzyme was 45 pmol min-' mg-' [lo] and that the M , was 520000 for a molecule containing two active sites [3,4]. Before all experiments, enzyme solutions were dialysed for 24 h against a 0.01 M potassium phosphate buffer, pH 7.4 (this buffer was employed for all the experiments described here). AdoCbl from Glaxo Research Ltd (Stoke Poges, Bucks, U.K.) was purified as described previously [ I ]
M. R. Hollaway, H. A. White, K. N. Joblin, A. W. Johnson, M. F. Lappert, and 0. C. Wallis
and concentrations of AdoCbl solutions were determined spectrophotometrically, employing a molar absorption coefficient at 341 nm of 1.28 x lo4 M-' cm-' [Ill. L-2-Aminopropanol from Aldrich Chemical Co. Ltd (Gillingham, Dorset, U.K.) and 2-aminoethanol, purchased from B.D.H. Ltd (Poole, Dorset, U.K.), were purified by distillation prior to use. Other chemicals were of the best commercial grade available. As the holoenzyme is photosensitive, all the experiments and manipulations involving coenzyme described here were carried out in a room illuminated with dim red light and all solutions of reactants were carefully degassed and then kept under an atmosphere of 02-free argon. Rapid-wavelength-scanning, stopped-flow, spectrophotometric experiments were carried out by using the apparatus described by Hollaway and White [2]. The apparatus allows two solutions of reactants to be mixed within 3 ms and then spectra to be recorded at the rate of 800 s-', each spectrum taking 1 ms, with a 0.25 ms gap between spectra. In the present experiments each spectrum covered the range of wavelength from 345 nm to 570 nm. The apparatus is a thermostatted, dual-beam instrument and the electronic circuitry gives an output voltage which corresponds to the difference in absorbance between the stopped-flow cuvette and the reference cuvette. After analogue to digital conversion the resulting spectra are stored in an ESL data-capture system (ESL Ltd, High Wycombe, Bucks, U.K.), which enables the storage of up to 32 spectra (each comprising 256 eight-bit words) at preselected times during the reaction. At the fastest rate of working, 32 spectra can be recorded in 40 ms, but for slower reactions appropriate circuitry enables the recording of every nth spectrum, where n = 2, 5, 10, 25, 50 or 100. The data-capture device can also be used in a signal-average mode and in the present work spectral changes in some of the slower reactions (i.e. half lives of seconds) were followed by summing 64 spectra in a single register. In this case, each averaged spectrum takes about 0.15 s to accumulate but the signalto-noise ratio is improved. After storage of spectra in the data capture system, they can be displayed on an oscilloscope screen, photographed and drawings made from the negatives by projection on to graph paper, using a photographic enlarger. Time courses at any required wavelength can also be displayed by using the 'point n' mode which causes the memory to output the ith points in each spectrum. The spectra shown in the figures were obtained by subtraction of the appropriate baselines. Calibration of wavelengths was achieved by identifying the positions of peaks given by neodymium oxide and holmium oxide filters.
145
An indication of the quality of the data presented here can be obtained by inspection of Fig. 1, which shows typical spectra obtained during the reaction with L-2-aminopropanol as substrate. The spectrum shown in Fig. 1A was recorded in 1 ms and after subtraction of the baseline gives the corrected spectrum (in Fig. 1 C, dotted lines). This spectrum is closely similar to the averaged spectrum obtained from the traces of Fig. 1 B, (see Fig. 1 C, solid line), although it has a poorer signal-to-noise ratio (compare Fig. 1 A and B). We have established that the spectra recorded by using the 'average' mode (64 spectra averaged) of the data-capture system are indistinguishable from spectra recorded by using a wavelength and absorbancecalibrated Cary 14 spectrophotometer and that individual spectra are accurate to within about 5 % at all wavelengths. The 8-bit resolution of the analogueto-digital conversion limits the precision of our measurements entirely with averaged spectra and to some extent with individual spectra. RESULTS Previous investigations of the transient kinetics of the reaction catalysed by ethanolamine ammonialyase [ l , 121 involved rapid mixing of apoenzyme (syringe 1) with a solution containing coenzyme and substrate (syringe 2). It seemed possible that these results could have been affected by a slow formation of holoenzyme so we have carried out two sets of experiments with both L-2-aminopropanol and 2-aminoethanol as substrates. (Catalytic centre activities are about 1 s - ' and 140 s - ' respectively with these substrates.) In the first series of experiments the enzyme solution was placed in one syringe of the rapid-wavelengthscanning, stopped-flow apparatus and a mixture of AdoCbl and substrate in the other syringe. In the second series, the enzyme and coenzyme solutions (syringe 1) were mixed with the substrate solution (syringe 2). It can be seen from the following results that this alteration in the order of addition had a profound effect on the rate of appearance of the cob(I1)alamin intermediate. PRE-STEADY-STATE SPECTRAL CHANGES WITH L-2-AMINOPROPANOL AS SUBSTRATE
Apoenzyme (Syringe I) Mixed with Coenzyme and Substrate (Syringe 2) We have presented elsewhere results from a study of this reaction [l], but further data are presented here as our recent experiments were conducted with a wider range of wavelength sweep (about 340- 560 nm) than that employed in the earlier investigation (400550 nm). This extended range gives access to the near-
146
Ethanolamine Ammonia-lyase: a Transient Kinetic Study
350
400
350
400
450 500 Wavelength (nrn)
550
(u
, " t 00:
SI
~
0 15
450
500
550
- c
u m
:
ultraviolet region where absorption bands can be useful in identifying certain cobalamin derivatives, e.g. cob(1)alamin has a strong absorption band at 386 nm. Fig. 2A shows spectra recorded 3,312 and 1250 ms after mixing enzyme (28 pM in active sites) with a solution containing 24 pM coenzyme (AdoCbl) and 4 mM L-2-aminopropanol. The spectrum observed after 3 ms reaction corresponds closely to that of unmodified coenzyme (AdoCbl) and the spectrum after 1.25 s is similar to the spectrum of cob(I1)alamin with absorbance maxima at 380 and 475 nm and shoulders at 460 and 520 nm (cc Fig.2A and C). However, in view of the close similarity between the spectra of 'base-on' and 'base-off Co(I1) cobalamins [15] in this region of the absorption spectrum, it is difficult to be unequivocal as to the exact identity of the Co(I1) species (see Discussion). The time courses for the absorbance changes during the reaction, at three different wavelengths, are identical within experimental error (Fig. 2B) and there are isosbestic points at 393 and 486 nm (Fig. 2A). Thus, the formation of cob(I1)alamin appears to involve a single step, although the possibility cannot be excluded that other intermediates are formed in stoichiometrically insignificant concentrations during the reaction. The value of the apparent first-order rate constant for the formation of the cob(I1)alamin intermediate, calculated from the logarithmic plots of Fig. 2B, is 4.7 s-', which is similar in magnitude to the value of 3 s-l obtained previously [ I ] and comparable to the value of 7 sC1 reported by Babior et al. [6] for the formation of a paramagnetic Co(I1) species during the reaction of ethanolamine ammonia-lyase with the same substrate and coenzyme. (The significance of these values is discussed below.)
n 010
b n
A Spectrophotometric Rapid Kinetic Study of Reactions Catalysed by Coenzyme-BIZ-Dependent Ethanolamine Ammonia-Lyase Michael R. HOLLAWAY, Hugh A. WlIITE, Keith N. JOBLIN, Alan W. JOHNSON. Michael F. LAPI'ERT, and 0. Caryl WALLIS Department of Biochemistry, University College London and School of Molecular Sciences, University of Sussex (Received January 11, 1977)
1. Spectral changes during the transient states and the steady states of the reactions catalysed by the adenosylcobalamin-dependent ethanolamine ammonia-lyase with L-2-aminopropanol and 2-aniinoethanol as substrates were followed by using a rapid-wavelength-scanning, stopped-flow spectrophotometer (800 spectra per second and each spectrum 340- 560 nm). 2. During the steady state of the reaction with the 'poor' substrate, L-2-aminopropanol (k,,, = 1 s-l), under conditions of V , the spectrum of the coenzyme closely resembled that of cob(I1)alaniin. With the 'good' substrate, 2-aniinoethanol (k,,, z 140 s - I ) , under the same conditions, the spectrum corresponded to a mixture of 58 0,; cob(I1)alamin and 42 o/, cob(I1I)alamin. The latter may be either unchanged adenosylcobalamin or a coenzyme derivative with a covalent bond between a substrate carbon and the cobalt atom. Thc magnitude of a reported kinetic deuterium isotope effect favours the second possibility. There was no evidence for the presence of a cob(1)alamin intermediate. 3. The transient phases of the reactions were dependent on the order of mixing of reactants. When enzyme was placed in one syringe of the stopped-flow apparatus and coenzyme and substrate in the other, the cob(1I)alamin intermediate formed slowly (first-order rate constants of about 5 s-' with L-2-aminopropanol as substrate and about 2 s-l with 2-aminoethanol). When the enzyme and coenzyme were placed in one syringe and the substrate in the other, the formation of the cob(1I)alamin occurred much more rapidly (90 s - ' with L-2-aminopropanol and greater than 300 s-' with 2-aminoethanol). It is concluded that binding of adenosylcobalamin to the enzyme is followed by a slow change in the conformation of the enzyme molecule to give a catalytically active species (there is no observable change in the spectrum of the coenzyme concomniitant with this process). The formation of active holoenzyme is slow on the time scale of subsequent catalytic steps when 2-aminoethanol is the substrate. 4. The spectra of the intermediates formed in the reactions with both substrates, the rates of their formation and their lifetimes allowed a detailed kinetic analysis of the reactions in terms of a threestep mechanism involving binding of substrate. cob(I1)alamin formation ( k + z step) and cob(I1)alamin breakdown ( k + 3step). I n catalysis the difference between the substrates is expressed mainly in the /i+, step (240 s-l and about 1 s-l respectively for 2-aminoethanol and L-2-aniinopropanol, the respective k + values being 336 s- and 90 s '). ~
We have continued our studies [ l ] on the mechanism of action of the adenosylcobalatnin-dependent ethanolamine ammonia-lyase from Clostridiian sp., as approached by the measurement of the spectra of reaction intermediates by stopped-flow, rapid-scanning spectrophotometry [2]. These experiments in-
volve mixing enzyme and substrate solutions in about 3 ins and then recording spectra over a 200-nm range at the rate of 800 s-'. Ethanolamine ammonia-lyase catalyses the formation of acetaldehyde and ammonia from 2-aminoethanol or of ammonia and propionaldehyde from L-2-aminopropanol, in reactions which each involve the of an amino group (which may be proThis paper is Part V of the series C o r n ~ ~ , ) ~ c , - B , ~ - ~ c , / ~ r , ~ c c / rmigration /rt tonated. Scheme 1 ) between substrate carbon atoms. Rf,ocriorls (for Part 1V see [9]). The enzyme molecule. molecular weight 520000, A h h ~ ~ i ~ t i AdoCbl. ~ i i s . adenosylcobalamin; ESR, electron spin contains two active sites which appear to function resonance. EKJWK,. Ethanolamine ammonia-lyast: (EC 4.3.1.7). independently [3,4].
144
Ethanolamine Ammonia-lyase: a Transient Kinetic Study HO R-CH,
I
+ I
NH,
CH,-CH,
+ I
HO
NH,
I
R-CH,
CH,-CH,
AdoCbl
1
PH
HO
$H-CH,
I
I
I
t
I I
I
/"" CH,-CH R-CH,
AH,
I N
R-CH,
HO R-CH,
+
NH,
r-CH2 I
I
Scheme 1. Mechanistic .scher~ir,/or. the r'ructioti ccitalysrtl hy eihanolanzinr urnnioriici-/)use with 2-aniinoeihanol as suhsirate R - CH2 - = 5'-adenosykobdkimin The catalytic steps of the reaction are believed to proceed by a mechanism which may (stippled arrows) or may not involve substrate-metal binding during rearrangement. A similar scheme can. in principle, be drawn for L-2-aminopropanol as substrate
In a previous communication [l], it was demonstrated that, with L-2-aminopropanol as substrate and AdoCbl [containing Co(III)] as cofactor, the characteristic spectrum of a Co(I1) form of the coenzyme [designated as cob(II)alamin, formerly as B I ~ , ]appeared during the enzymic reaction in a process with a first-order rate constant of about 3 s-l. As the catalytic centre activity for ethanolamine ammonia-lyase with L-2-aminopropanol as substrate is between 1 and 2 s-l [ 5 ] , the rate of formation of the cob(I1)alamin species was consistent with it being an intermediate on the catalytic pathway. This observation was in agreement with previous freeze-quenching ESR spectroscopic studies with the same substrate and cofactor [6], which indicated that a paramagnetic Co(I1) species formed at a kinetically competent rate after mixing enzyme with AdoCbl and L-2-aminopropanol. Given that the Co(I1) species of the coenzyme forms at a kinetically competent rate during catalysis, then a radical rearrangement mechanism seems likely for the reaction catalysed by ethanolamine ammonia-lyase (see Scheme l ) , at least with L-2aminopropanol as substrate, although the intervention of a carbonium ion rearrangement cannot be finally excluded on the evidence available at present. The catalytic centre activity for ethanolamine ammonia-lyase with 2-aminoethanol as substrate is about 140 s - ' [4] and this is 100-fold greater than that with L-2-aminopropanol, so care must be exercised in extrapolating conclusions reached by consideration of results obtained by using the 'poor' substrate to the reaction with the 'normal' substrate. Nevertheless,
it has been shown [l,71, that a cob(I1)alamin species of the coenzyme is present during the reaction catalysed by ethanolamine ammonia-lyase with 2-aminoethanol as substrate. However, there has been no demonstration of the kinetic competence of the cob(I1)alamin species as an intermediate in this reaction and part of the present study was directed towards the investigation of this point. In addition, results are presented which demonstrate that ethanolamine ammonia-lyase undergoes a structural change on binding coenzyme, to give a catalytically active species, and that, with both L-2-aminopropanol and 2-aminoethanol as substrate, this change is much slower than subsequent turnovers. MATERIALS AND METHODS Ethanolamine ammonia-lyase was purified from cultures of Clostridium sp. by the method of Kaplan and Stadtman [8] with the modifications described by Joblin et ul. [9]. Bound cobalamin derivatives were removed from the enzyme by the procedures described in these papers. Molar concentrations of enzyme active sites in reactant solutions were calculated from measurements of enzyme activity, assuming that the specific activity of the enzyme was 45 pmol min-' mg-' [lo] and that the M , was 520000 for a molecule containing two active sites [3,4]. Before all experiments, enzyme solutions were dialysed for 24 h against a 0.01 M potassium phosphate buffer, pH 7.4 (this buffer was employed for all the experiments described here). AdoCbl from Glaxo Research Ltd (Stoke Poges, Bucks, U.K.) was purified as described previously [ I ]
M. R. Hollaway, H. A. White, K. N. Joblin, A. W. Johnson, M. F. Lappert, and 0. C. Wallis
and concentrations of AdoCbl solutions were determined spectrophotometrically, employing a molar absorption coefficient at 341 nm of 1.28 x lo4 M-' cm-' [Ill. L-2-Aminopropanol from Aldrich Chemical Co. Ltd (Gillingham, Dorset, U.K.) and 2-aminoethanol, purchased from B.D.H. Ltd (Poole, Dorset, U.K.), were purified by distillation prior to use. Other chemicals were of the best commercial grade available. As the holoenzyme is photosensitive, all the experiments and manipulations involving coenzyme described here were carried out in a room illuminated with dim red light and all solutions of reactants were carefully degassed and then kept under an atmosphere of 02-free argon. Rapid-wavelength-scanning, stopped-flow, spectrophotometric experiments were carried out by using the apparatus described by Hollaway and White [2]. The apparatus allows two solutions of reactants to be mixed within 3 ms and then spectra to be recorded at the rate of 800 s-', each spectrum taking 1 ms, with a 0.25 ms gap between spectra. In the present experiments each spectrum covered the range of wavelength from 345 nm to 570 nm. The apparatus is a thermostatted, dual-beam instrument and the electronic circuitry gives an output voltage which corresponds to the difference in absorbance between the stopped-flow cuvette and the reference cuvette. After analogue to digital conversion the resulting spectra are stored in an ESL data-capture system (ESL Ltd, High Wycombe, Bucks, U.K.), which enables the storage of up to 32 spectra (each comprising 256 eight-bit words) at preselected times during the reaction. At the fastest rate of working, 32 spectra can be recorded in 40 ms, but for slower reactions appropriate circuitry enables the recording of every nth spectrum, where n = 2, 5, 10, 25, 50 or 100. The data-capture device can also be used in a signal-average mode and in the present work spectral changes in some of the slower reactions (i.e. half lives of seconds) were followed by summing 64 spectra in a single register. In this case, each averaged spectrum takes about 0.15 s to accumulate but the signalto-noise ratio is improved. After storage of spectra in the data capture system, they can be displayed on an oscilloscope screen, photographed and drawings made from the negatives by projection on to graph paper, using a photographic enlarger. Time courses at any required wavelength can also be displayed by using the 'point n' mode which causes the memory to output the ith points in each spectrum. The spectra shown in the figures were obtained by subtraction of the appropriate baselines. Calibration of wavelengths was achieved by identifying the positions of peaks given by neodymium oxide and holmium oxide filters.
145
An indication of the quality of the data presented here can be obtained by inspection of Fig. 1, which shows typical spectra obtained during the reaction with L-2-aminopropanol as substrate. The spectrum shown in Fig. 1A was recorded in 1 ms and after subtraction of the baseline gives the corrected spectrum (in Fig. 1 C, dotted lines). This spectrum is closely similar to the averaged spectrum obtained from the traces of Fig. 1 B, (see Fig. 1 C, solid line), although it has a poorer signal-to-noise ratio (compare Fig. 1 A and B). We have established that the spectra recorded by using the 'average' mode (64 spectra averaged) of the data-capture system are indistinguishable from spectra recorded by using a wavelength and absorbancecalibrated Cary 14 spectrophotometer and that individual spectra are accurate to within about 5 % at all wavelengths. The 8-bit resolution of the analogueto-digital conversion limits the precision of our measurements entirely with averaged spectra and to some extent with individual spectra. RESULTS Previous investigations of the transient kinetics of the reaction catalysed by ethanolamine ammonialyase [ l , 121 involved rapid mixing of apoenzyme (syringe 1) with a solution containing coenzyme and substrate (syringe 2). It seemed possible that these results could have been affected by a slow formation of holoenzyme so we have carried out two sets of experiments with both L-2-aminopropanol and 2-aminoethanol as substrates. (Catalytic centre activities are about 1 s - ' and 140 s - ' respectively with these substrates.) In the first series of experiments the enzyme solution was placed in one syringe of the rapid-wavelengthscanning, stopped-flow apparatus and a mixture of AdoCbl and substrate in the other syringe. In the second series, the enzyme and coenzyme solutions (syringe 1) were mixed with the substrate solution (syringe 2). It can be seen from the following results that this alteration in the order of addition had a profound effect on the rate of appearance of the cob(I1)alamin intermediate. PRE-STEADY-STATE SPECTRAL CHANGES WITH L-2-AMINOPROPANOL AS SUBSTRATE
Apoenzyme (Syringe I) Mixed with Coenzyme and Substrate (Syringe 2) We have presented elsewhere results from a study of this reaction [l], but further data are presented here as our recent experiments were conducted with a wider range of wavelength sweep (about 340- 560 nm) than that employed in the earlier investigation (400550 nm). This extended range gives access to the near-
146
Ethanolamine Ammonia-lyase: a Transient Kinetic Study
350
400
350
400
450 500 Wavelength (nrn)
550
(u
, " t 00:
SI
~
0 15
450
500
550
- c
u m
:
ultraviolet region where absorption bands can be useful in identifying certain cobalamin derivatives, e.g. cob(1)alamin has a strong absorption band at 386 nm. Fig. 2A shows spectra recorded 3,312 and 1250 ms after mixing enzyme (28 pM in active sites) with a solution containing 24 pM coenzyme (AdoCbl) and 4 mM L-2-aminopropanol. The spectrum observed after 3 ms reaction corresponds closely to that of unmodified coenzyme (AdoCbl) and the spectrum after 1.25 s is similar to the spectrum of cob(I1)alamin with absorbance maxima at 380 and 475 nm and shoulders at 460 and 520 nm (cc Fig.2A and C). However, in view of the close similarity between the spectra of 'base-on' and 'base-off Co(I1) cobalamins [15] in this region of the absorption spectrum, it is difficult to be unequivocal as to the exact identity of the Co(I1) species (see Discussion). The time courses for the absorbance changes during the reaction, at three different wavelengths, are identical within experimental error (Fig. 2B) and there are isosbestic points at 393 and 486 nm (Fig. 2A). Thus, the formation of cob(I1)alamin appears to involve a single step, although the possibility cannot be excluded that other intermediates are formed in stoichiometrically insignificant concentrations during the reaction. The value of the apparent first-order rate constant for the formation of the cob(I1)alamin intermediate, calculated from the logarithmic plots of Fig. 2B, is 4.7 s-', which is similar in magnitude to the value of 3 s-l obtained previously [ I ] and comparable to the value of 7 sC1 reported by Babior et al. [6] for the formation of a paramagnetic Co(I1) species during the reaction of ethanolamine ammonia-lyase with the same substrate and coenzyme. (The significance of these values is discussed below.)
n 010
b n
