Fur J Biochcm 111, 177-188 (1980)

C, by FEBS 1980

The Number of Functional Active Sites per Molecule

of the Adenosylcobalamin-Dependent Enzyme,
Ethanolamine Ammonia-lyase, as Determined by a Kinetic Method
Michael R . HOLLAWAY, Alan W. JOHNSON, Michael F. LAPPERT, and 0 . Caryl WALLIS
Department of Biochemistry, University College, London, and
School of Molecular Sciences, University of Sussex, Brighton
(Received January 2S/May 9, 1980)

1. A kinetic approach to the determination of the number of functional active sites per molecule
of the adenosylcobalamin-dependent enzyme, ethanolamine ammonia-lyase, is described.
2. Time courses for formation and breakdown of a cob(I1)alamiii intermediate during reaction
of the enzyme, fully saturated with adenosylcobalamin, with L-2-aminopropanol as substrate, were
followed using a stopped-flow spectrophotometer under two conditions : (a) enzyme concentration
much greater than that of substrate, (b) substrate concentration much greater than that of enzyme.
3. Results were analysed in terms of a three-step mechanism involving binding of substrate
(k+1 step), cob(I1)alainin formation ( k + 2 step) and cob(I1)alamin breakdown (k+3 step). The kinetic
scheme was shown to be sufficient to account for the observed time courses and rate constants
of 80 sK1 (k+z) and 1.5 s-' ( k + 3 )were determined.
4. The number of active sites per enzyme molecule ( n ) was calculated from the kinetic data in
three ways : (a) calculation from amplitude of absorbance measurements, (b) calculation from measurements of the values of rate constants and (c) analysis by computation of the kinetic data using
the computer program FACSIMILE. A value for TI close to 6 was calculated by each of these
methods. This value is in disagreement with the literature value of about two sites per molecule
but is consistent with the I& subunit structure of the enzyme.
5. Kinetic analysis of data from experiments in which the adenosylcobalamin concentration was
varied while substrate and enzyme concentrations remained constant showed that all the active
sites function with identical rate constants.
6. The principle and mathematical basis of the kinetic method for determining the value of II is
given as an Appendix.

Ethanolamine ammonia-lyase is an enzyme dependent on adenosylcobalamin (AdoCbl) which catalyses the formation of acetaldehyde and ammonia
from 2-aminoethanol, or of propionaldehyde and ammonia from 2-aminopropanol, in reactions each involving the migration of an amino group between
adjacent substrate carbon atoms [l] as shown in
Scheme 1.
The molecular weight of the enzyme is about
520000 [2] and it has been reported that there are
two active sites per molecule [3 - 51. Recent work [6]
has shown that the enzyme molecule probably contains equal numbers of two types of subunit (I and II),
of M , 51 000 and 36000, indicating a subunit structure for the native enzyme of 16116.
Abbreviations. AdoCbl, adenosylcobalamin ; C D , circular dichroism.
Enzyme. Ethanolamine ammonia-lyase (EC 4.3.1.7).

In previous communications [7,8] we have reported kinetic results using rapid-scanning, stoppedflow spectrophotometry and have demonstrated the
existence of kinetically competent Co(I1) intermediates
during the reaction of ethanolamine ammonia-lyase
with 2-aminoethanol or with L-2-aminopropanol.
From these data a detailed kinetic analysis was derived
in terms of a three-step mechanism (Scheme 2, see
below) involving binding of substrate, formation of
the cob(1I)alamin intermediate and its breakdown.
In this paper we describe the results of a kinetic
approach to the determination of the number of
functional active sites in the ethanolamine ammonialyase molecule. The method used employs two types
of kinetic information : one from experiments where
the concentration of enzyme is much greater than
that of substrate and the other where the concentration
of substrate is much greater than that of the enzyme
and the K , value for the substrate. In each case time

178

Number of Active Sites in Ethanolamine Ammonia-lyase

HO
R-CH?

kH,

HO

iH,

HO

R-CH,

AH,

CH,-CH-CH,

-\,,,,,,\

W
i
I
I
)

AdoCbl

OH

R-CHz

7"

CHZ-CHZ-CH

R-CH,

CH,-$H

OH
/
-CH

+
NH,

HO

R-CH,

CH-CH-CH,

\+

\+

NH3

NH,

CH3

CH,-CH

<*C-R

HO

OH

NH,

:-&,
CH-CH- CH,

------

Scheme 1. Mecliunisnzfor the reaction catulysed h j cthnnolumi~rcumnzoniu-/yu.sr with I.-2-umino~~royuriol

us suhatratr. R CH2 = 5'-adenosyl.
The catalytic steps of' the reaction are believed to proceed by a mechanism which may o r may not involve substrate-metal binding
during the rearrangement. One carbon atom is shown in bold typc t o illustrate thc rearrangement
-

courses of formation and breakdown of the cob(I1)alamin intermediate were followed by using a stoppedflow spectrophotorneter. The results indicate that
there are six functional active sites per molecule of
enzyme.

MATERIALS A ND METHODS
Ethanolamine ammonia-lyase was purified from
cultures of Clostridiuin sp. and bound cobalamins removed by the method of Kaplan and Stadtman [9]
with the modifications described by Joblin et al. [lo].
More reliable growth of clostridia and higher yields
of enzyme were obtained with a low sparging rate
with 957; Nz/5';/, COz during enzyme induction.
2-Aminoethanol was removed from the enzyme by
dialysis for 24 h against 0.01 M potassium phosphate
buffer, pH 7.4, and all experiments were carried out
using the same buffer.
Initially, enzyme concentration was measured by
dry weight determinations on solutions of purified
enzyme and confirmed by microanalysis for carbon
and nitrogen content. Routinely the concentration
of enzyme solutions was compared with this standard
by microanalysis for carbon and nitrogen, measurement of the absorbance at 280 nm and protein estimation by the biuret and Lowry procedures [l 1,121 using

bovine serum albumin as a standard. It was found

that the protein concentrations of enzyme solutions
determined by the biuret or Lowry procedures were
very close and did not differ by the factor of 1.54
described by Kaplan and Stadtman [9]. The results
of the Lowry determinations required multiplication
by a factor of 0.89 to give the same concentrations
as those determined from the dry weight. Molar concentrations of enzyme solutions were calculated from
the protein concentrations by using a molecular weight
of 520000 [2] for the native enzyme.
Time courses of formation and breakdown of the
cob(I1)alamin intermediate were followed at 528 nm
by using a Durrum-Gibson stopped-flow apparatus
(Durrum Inst. Corp., Palo Alto, CA, USA) equipped
with a logarithmic amplifier. The pathlength of the
observation cuvette was 1.7 cm. Negatives of photographs of the time courses were used to project the
traces onto graph paper for drawing by hand. In all
experiments enzyme and AdoCbl were placed in one
syringe and the substrate, L-2-aminopropanol, in the
other syringe. These conditions ensured that the active
enzyme-AdoCbl complex was present before addition
of substrate [8].
Computer fitting of kinetic data to Scheme 2 was
carried out by using the program FACSIMILE [13].
This program determines the 'best-fit' values of rate
constants given a digitized time course of a reaction.

M. R. Hollaway, A. W. Johnson, M. F. Lappert, and 0. C. Wallis

179

The optimization routine involves minimization of

the normalized residual sum of squares given by

changed if this rate constant is included (see Appendix).

The determination of the number of active sites
per molecule based on measurements of kinetic
parameters and reaction amplitudes (a) with the concentration of enzyme active sites much greater than
that of substrate and (b) with substrate concentration
much greater than that of active site, depends on a
fairly complete knowledge of the kinetics of the
particular reaction under study (see Appendix). In
the following text we present first the results of the
kinetic investigations of the formation and breakdown of the cob(I1)alamin intermediate which were
carried out to test the conformity to Scheme 2, and
then the results of the evaluation of the number of
active sites per enzyme molecule.

where j identifies the data points for the ith time

course, m is the number of time courses fitted and n
the number of points for each time course. The
residuals R i j are given by the difference between the
observed and calculated values of the independent
variable, ( v i j - uij), divided by the largest value ofthat
variable, r , divided by the standard error of the time
course, rs,, which was taken as 1 % for the data
presented here.
AdoCbl from Glaxo Research Ltd (Stoke Poges,
Bucks, UK) was purified and its concentration determined as described previously [7,8]. All manipulations
involving AdoCbl or holoenzyme were carried out
in a room illuminated with a dim red light.
L-2-Aminopropanol was obtained from Aldrich
Chemical Co Ltd (Gillingham, Dorset, UK) and was
purified by distillation before use.
Bovine serum albumin was obtained from Sigma
Chemical Co (Poole, Dorset, UK). Concentrations of
solutions were determined from the absorbance at
279 nm on a Cary 14 spectrophotometer usingA;?,,,
= 6.60 [14]. This value was checked by carrying out
dry weight determinations on solutions of bovine
serum albumin of known A2,9 and the value so obtained of 6.60 & 0.21 confirms the accuracy of the
dry weight measurements made in this study.
All other chemicals were of the best commercial
grade available.

RESULTS
Time courses of formation and breakdown of the
cob(I1)alamin intermediate were analysed on the basis
of the kinetic mechanism given in Scheme 2 [ 8 ] :
EC+S

S KSGERPL+EC+P
1%

Scheme 2
where EC represents the enzyme-AdoCbl complcx ;
S, L-2-aminopropanol ; ERP, the cob(1I)alamin intermediate (or collection of intermediates) in which the
coenzyme is in the form of cob(1I)alamin and P the
presumed primary product : l-aminopropanol, which
dissociates to propionaldehyde and ammonia rapidly
on the time scale of the experiments reported here
(cf. [ 1 5 ] ) . The first step in the reaction can be taken
to be a pseudo-equilibrium process [8], with an apparent dissociation constant, K, = k 1 / k + I . The consideration of a k-2 step in Scheme 2 is omitted as
the form of the rate laws and the argument relating
to the determination of the stoichionietry of the interaction between enzyme and substrate remains un-

KINETICS OF THE REACTION

WITH L-2-AMINOPROPANOL AS SUBSTRATE

Enzyme Concentration Much Greater than that

qf Suhstrute ( [ E l o 9 [ S ] )
Under the condition [ E l 0 % [S]O, and assuming
that Scheme 2 applies, the concentration of the cob(I1)alamin intermediate, [ERP], at any time t is given by
Eqn (1) :

where

and [El0 and [S]Oare the initial enzyme and substrate

concentrations. In the treatment used to derive this
equation it has been assumed that the formation of
ECS (i.e. the enzyme-substrate complex) can be regarded as a pseudo-equilibrium process (see Appendix). This requires that
k + 2 ,and this condition
is fulfilled as it will be shown that the order of
magnitude of k-1 is lo3 SKIor greater whereas the
value of k + z is about 80 s-'.
Fig1 shows photographs or the time courses of
formation and decay of the cob(I1)alamin intermediate. The signal-to-noise ratio in the time course
for its formation (Fig. 1 A) is poor as a result of a small
and fairly rapid absorbance change and so it is only
possible to make an estimate for the rate constant
k ; z of about 40 s-'. Other time courses with different
concentrations of enzyme in the range 1.2-3.5 pM
gave rate constants spanning 25 - 50 s-' but the signal :
noise ratio in the traces was not sufficiently good to
enable the determination of the dependence of k;2
on the concentration of enzyme. Nevertheless in all
cases the value of the rate constant for the formation
of the cob(I1)alaniin intermediate was more than
16 times that for its breakdown.

Number of Active Sites in Ethaiiolaminc Ammonia-lqase

180

Fig. 1 . Time course of ,fijrmulion ctnd hr~.aXrloisnof ihe <.oh(ll)uluniin iitiermc4Lirc isit17 .suhsiruit, w i i l i'nirutioii I C ~ S A ihuir rhur of' mrviric,.
Pholographs of time co~irsesfor formation ( A ) and breakdown (B) of the intcrinediate alter mixing en7ymc ( 7 p M ) + AdoCbl (126 p h l ) i n
syringe 1 with 1.-2-aminopropanol (4.36 p M ) in syringe 2. Concentrations in the reaction mixture were one half of those in the ,>ringes
as there is 1 : 1 dilution on mixing. The buffer was 0.01 M potassium phosphate, pH 7.4 in this and all the reactions described here. and
the temperature was 27 "C

Fig. 1B shows a time course for breakdown of the

cob(II)alamin intermcdiate. The value of the firstorder rate constant for this process was calculated as
1.5 k 0.1 s-' and was found to be independent of
enzyme concentration in the range from 3.12 pM to
below 1.28 pM. These results, taken in conjunction
with the observation that the absorption spectrum of
the cob(I1)alamin intermediate recorded during the
steady state of the ethanolamine ammonia-lyasecatalysed reaction with L-2-aminopropanol as substrate corresponds to that expected for a nearly
stoichiometric conversion of coenzyme to cob(I1)alamin [8]. indicate that the value of about 1.5 s-l
for breakdown of the cob(I1)alamin intermediate can
be assigned to the k + 3 step in Scheme 2. Correspondingly the value of about 40 s-' can be assigned
to the rate constant for the k + 2 step.
It can be shown (see Appendix) that the maximum
value of the concentration of intermediate ([ERP],,,)
reached in time courses such as those shown in Fig. 1
is given by Eqn (2) :
(2)
By substitution of the values of k + Zand k+3 given
above in Eqn (2), it can be calculated that about 90
ofthe substratewillbe in the form ofthe cob(I1)alamin
intermediate at the turning point in the traces of Fig. 1.
Suhstrute Concentration Much Greater thun that
of'Enzyme ( ( S ] OB [ E l o )
In this case an analytical treatment of Scheme 2
gives the following expression for the concentration
of active sites in the form of cob(I1)alamin intermediate at any time t during the approach to the
steady state (see Appendix).

where n is the number of active sites per molecule

and kY2 = k + z / ( l K , / [ S ] o ) .

This corresponds to' a pseudo-first-order process

characterized by a rate constant, kTbs, given by

+ /<+A).
and amplitude ,4 = k ' ; ] . II . [EIoi(k';,
In the derivation of Eqn (3) it is assumed that the substrate concentration changes by a negligible amount
during the approach to the steady state, a condition
that holds to within 10% in the experiments described here.
From Eqn (4) it can be seen that kPbs should be a
hyperbolic function of substrate concentration and
in the limit, when [S]O9 K,, should approach the
sum of the values of the rate constants for formation
and breakdown of the cob(1I)alamin intermediate,
i.e. (k+2 k + 3 ) , see Scheme 2. A value for K, can be
calculated from the dependence of kpbSon [S]" using
conventional graphical procedures.
Fig.2A shows a photograph of the time course
of formation of the cob(I1)alamin intermediate in a
reaction mixture containing 1.2 yM enzyme and 87 yM
substrate. The same photograph shows the time course
for the overall reaction, from which it can be seen
that the concentration of the intermediate changes
little for about 10 s after the steady-state value is
reached but then falls rapidly, approaching zero at
14 s. For a given substrate concentration the observed
rate constant for formation of the cob(I1)alamin
intermediate was independent of enzyme concentration as long as the condition [S]o 9 [E]o held. However
the duration of the steady-state period decreased with
increasing enzyme concentration in the way expected
for the cob(1I)alamin species acting as an intermediate
on the catalytic pathway. A computer-based analysis
of time courses, such as those of Fig.2, will be given
below.
The value of the first-order rate constant calculated
from the time course for formation of the cob(I1)alamin intermediate shown in Fig.2 is 29 s P 1 but
higher concentrations of substrate gave greater kYbs

M . R . I-lolla~ay,A . W. Johnson, M . F. Lappert. and 0. C. Wallis

3
Time ( s )

Fig. 2. Time course offortnation and breakdown qf the coh(1l)ulamin intermediate and of product formation with suhstvate concentrations
muck greater than ofenzyme. (A, B) Photographs of time courses for formation and breakdown of the intermediate when 1.76 pM enzyme
+ 126 pM AdoCbl (syringe 1) were mixed with 274 pM L-2-aminopropanol (syringe 2). Enzyme concentration was 0.88 pM and substrate
concentration 87 pM in reaction mixtures. The remperature was 27C. (C) Time course of product formation in the reaction shown
in (B) calculated as described in the text

values, eventually approaching 80

10 s- (see
Fig.2B). Thus ( k + z+ k + 3 ) is about 80 spl and, as
k + 3 is only 1.5 spl (see previous section), k + z is also
close to 80 s - l . The concentration dependence of kTbS
in these reactions is consistent with a value for K, of
about 100 pM.
The overall time course for formation and breakdown of the cob(I1)alamin intermediate shown in
Fig.2A and B can be integrated to give the progress
curve for formation of products by the following
procedure. As the amount of product formed at any
time t , [PI,, is given by Eqn (5):

where E~~~ and EEC are the molar absorption coefficients

for species ERP and EC respectively and A, is the
absorbance at any time t. For complete reaction
( t = m ) , [PImwill correspond in value to [S]o so that:

where I , is the total area under the absorbanceitime

curve. Dividing Eqn ( 5 ) by Eqn (6) we have:

(7)
so that the required time course can be obtained by
measuring the areas under the absorbanceltime curve
at different times and multiplying these values by the
factor [Slo/Z,.
The result of effecting this transform on the
overall absorbance versus time profile of Fig. 2 A is
given in Fig. 2C. Higher substrate concentrations
gave the same constant slope of the curve of 7.9
x
M s- (i.e. the extent of formation of the
cob(I1)alamin intermediate cannot be made greater
than that of Fig.2A by increasing [S]O).So it can be
concluded that with a substrate concentration of
87 pM the steady-state rate corresponds essentially to
the maximum velocity (V) and we can then write:
V = kcat. ~1 . [El0

(8)

where k,,, is the catalytic centre activity of the enzyme

and n the number of active sites per molecule, which
for present purposes are presumed identical and
independent. The application of Eqn (8) to the deter-

1x2

Number of Active Sites in Ethanolamine Ammonia-lyase

mination of the number of active sites per enzyme

molecule is given in the following section.
TREATMENT OF THE KINETIC RESULTS
TO DETERMINE THE NUMBER O F ACTIVE SITES
PER MOLECULE OF ETHANOLAMINE AMMONIA-LYASE

The foregoing treatment involves the use of analytical solutions of differential equations arising from
Scheme 2, linearised by assumptions such as [El0 9 [S]O
or [Sl0 % [El0 and further simplified by the assumption
that the binding of substrate to the enzyme is a
pseudo-equilibrium process. (Justification of the pseudo-equilibrium binding of substrate to the enzyme
AdoCbl complex will be given below.)
Notwithstanding these assumptions a fairly clear
picture emerges of the overall enzymic reaction.
Following the formation of the ECS complex, which
occurs without any observable spectral change, conversion to the cob(I1)alamin-containing intermediate
takes place in a first-order process with a rate constant
of about 80 s-'. This species then breaks down to the
enzyme-AdoCbl complex and products in a process
with a first-order rate constant of about 1.5 s-'. The
assignment of the rate constants of 80 s-' and 1.5 s-l
to the formation and breakdown respectively of the
cob(I1)alamin intermediate rather than the converse,
which would be consistent with Eqns (1) and (3), is
based on the virtually stoichiometric conversion of
the enzyme-AdoCbl-substrate complex to the cob(I1)alamin intermediate (see above). Given a kinetic
scheme sufficient to account for the observed time
courses, the number of active sites per molecule then
can be calculated in the following three ways two of
which are independent, relying on measurements of
amplitudes and of rates respectively.
Culculation of n from Amplitude
of Absorbance Measurements
From the data of Fig. 1, 2.18 pM substrate and
3.5 pM enzyme gives a maximum absorbance change
of 0.01 15. With the same concentration of substrate
but a higher enzyme concentration there was not a
significant increase in this maximum observed extent
of cob(I1)alamin intermediate formation. In the limit
where the rate constant for breakdown of the cob(I1)alamin intermediate is insignificant in value compared
with that for its formation, the concentration of the
intermediate would correspond to that of substrate
and a value for ( E E ~ ~EEC) would be obtained of
about 3120 M-l . cm-'. However, substitution of the
known values for k+z and k+3 in Eqn (2) shows that
only about 90% of the substrate would be present
as the cob(1I)alamin intermediate at the maximum in
the change of absorbance, so the value (cERP - EEC)
requires correction to a value of about 3470 M-' . cm-'.

Fig. 3. A typical computerised unulysis of the reaction catulysrd by

etlrunoluwiine ammomu-lyuse with L-2-uminopropanof us substrute.
The reaction was carried out as described for Fig. 1 with the following concentrations of enzyme and substrate in the reaction
mixture: (a) 3.5 pM enzyme, 2.18 pM substrate; (b) 0.88 pM enzyme, 87 pM substrate. The computer fit, obtained using program
FACSIMILE, is given by the solid line and the experimental data
by the circles. Note that the abscissa is not linear in time. The
residual sum squares of the fit of experimental data to the optimized curve is 1.52 x lo3 for 174 data points. xz for the fit was 9.0
for each degree of freedom

The value of ( E E R P - CEC), calculated on the basis

of a known substrate concentration, can then be used
to calculate the concentration of the cob(I1)alamin
intermediate during the steady state of the reactions
where [El0 Q [Sl0. For example in Fig.2B the maximum absorbance difference is 0.032 for a 1.7-cm pathlength so [ERP],, = 5.4 pM [see Eqn (4)]. As the concentration of enzyme in this experiment was 0.88 pM,
it can be calculated after substitution in Eqn (4) that
there are 6.3 active sites per enzyme molecule.
Calcuhtion of n from Measurement

of the Values of Rate Constants

From the integrated time course of Fig. 2C it was
calculated that V = 7.9 x
M s-'. As the enzyme
concentration was 0.88 pM the value of n . k,,, is
9.05 s-' by substitution in Eqn (9).
As k,,, = k + z .k+3/(kf2+ k+3), with k+2 = 80 s-'
and k + 3 = 1.5 s-', then k,,, = 1.47 s-', so giving
n = 6.1 active sites per molecule, in excellent agreement with the value of 6.3 obtained from the amplitude measurements.
Analysis by Computation
ofthe Kinetic Data Using the Program FACSIMILE
Time courses of formation and breakdown of the
cob(I1)alamin intermediate such as those shown in
Fig. 1 and 2 were digitized and fitted to Scheme 2
using the program FACSIMILE developed by Chance
et al. [13]. A typical fit of the data is shown in Fig.3
and the best-fit values of the rate constants for four
time courses are as follows: k + 1 = 1.6 k 0.7 x 10' M-'
s-'; k-1 = 8.7 3 . 7 l~o 3 S - ' ; k + 2 = 78 15 s - ' ;
k+3 = 1.55 5 0.35 S - ' ; (EERP - E E C ) = 3820 & 180
M-' cm-' and n = 5.4 & 0.3.

M. R. Hollaway, A. W. Johnson, M . F. Lappert, and 0. C . Wallis

Inspection of Fig. 3 shows that there is a small but

systematic deviation of the observed points from the
theoretical curve and this could arise from the following sources: (a) slight loss of activity of the enzyme during the 12 s of the time courses with 87 pM
substrate; (this represents a real probability as hydroxycobalamin is formed with attendant loss of
activity when much greater concentrations of substrate are employed, see for example [8,16]; (b) accumulation of inhibitory products ; (c) inadequacy of
Scheme 2 to describe the mechanism; (d) functional
non-equivalence of active sites in the multimeric
enzyme.
It is possible that Scheme 2 is not sufficient to
describe the reaction but at the level of the present
experiments and considering the interference that
could arise from the sources (a) and (b), it can be
regarded as a reasonable basis for the conclusion
that n = 6 inasmuch that treatment of the results
according to kinetic mechanisms more complex than
Scheme 2 would not give lower values of n (see
Appendix).
A test for non-equivalence in the active sites was
devised and is described in the following section.

183

-0.01

2
p
a
-0.02

-0.03

10

q
n
I.

.
-

0.5

20
Time ( s )

30

lil :

L
,

0.2

2-

,.-.

10.1

Active Site Equivalence in the Reaction

with L-2-Aminopropanolas Substrate
0

Time courses of formation and breakdown of the

cob( 1I)alamin intermediate were recorded in reaction
mixtures containing fixed concentrations of enzyme
([El0 G [S]O)and substrate ([Sl0 9 K,) but different
concentrations of AdoCbl ranging from about one
sixth to double the concentration of available sites
(Fig.4A and B respectively). Results from these
experiments (Fig4 B) show that the concentration of
the cob(I1)alamin intermediate during the steady
state increases towards a saturating value with increasing coenzyme concentration.
The area under an absorbance/time curve of the
type shown in Fig.4A is given by I,, as defined in
Eqn (6), so:

00

10

IAdoCbll (pM)

(9)

Fig. 4. Equivalence of active sites in the etlzanolamine ammonia-lyasr

molecule. (A) Time courses obtained on mixing ( - - - - - ) 1.8 pM enzyme and 4.96 1M AdoCbl (syringe 1) with 100 pM L-2-amino1.8 pM enzyme and 24.8 pM AdoCbl
propanol (syringe 2); (--)
(syringe 1) with 100 pM L-2-aminopropanol. The temperature was
24.5 C and the other conditions were as for Fig. 1. (B) Dependence
of the maximurn steady state amplitude and area under the absorbancejtime course on the concentration of AdoCbl. (0)The relative maximum amplitudes of absorbance time curves such as those
shown in (A) at different concentrations of AdoCbl. The value of
d A 5 2 8 / d A , was calculated by division of the observed maximum
amplitude by the change in absorbance obtained by extrapolation
of the curve to infinite concentration of AdoCbl. ( 0 ) The total
area under the absorbance/time course (in arbitrary units) at given
concentrations of AdoCbl. The solid line is drawn for an apparent
dissociation constant of Kd = 0.4 pM and 5.4 AdoCbl binding
sites per molecule

where A c = ( E E ~ ~cEC).
For a given substrate concentration the value of
I, will depend on the value of k + 3 . If some active
sites in a molecule of enzyme are fast and others
slow then the value of k + 3 will vary with the number
of coenzyme molecules supplied per molecule of
enzyme, so giving a variation of the value of I , with
the degree of saturation of active sites on the enzyme
molecule. As the area under the absorbance versus
time curves was constant, (see Fig.4B, solid circles),
it can be concluded that the rate of transformation

of substrate per functioning active site is independent

of the number of AdoCbl molecules bound to a
given molecule of enzyme.
The data of Fig.4 can be analysed on the basis of
a phenomenological model in which the maximum
extent of formation of the cob(I1)alamin intermediate
in the steady state corresponds to that expected for a
simple equilibrium process between it and the enzymeAdoCbl complex. Of course, the observed apparent

Number of Active Sites in Ethanolamine Ammonia-lyasc

184

dissociation constant will not be a simple equilibrium

constant but will correspond to an apparent k, value
for the coenzyme. The value calculated for this
parameter from the data of Fig.4B was about 0.4pM.

DISCUSSION
Thc prescnt work has led to the conclusion that
there are six active sites per molecule of ethanolamine
ammonia-lyase and that these sites function at very
closely similar, if not identical, rates. The approach
described here for the determination of the number
of active sites per enzyme molecule depends on the
condition that all the sites function with identical
rate constants. The data of Fig.4 show that this
appears to be the case with ethanolamine ammonialyase acting on L-2-aminopropanol as substrate. It
should be noted that it is possible to exclude the interpretation that there is strong positive cooperativity
between AdoCbl binding sites such that non-saturating
concentrations of the coenzyme give a population of
enzyme molecules comprising only fully liganded and
coenzyme-free species. Given that there was initially
a homogeneous population of enzyme molecules, this
would give rise to a binding curve approximating two
intersecting straight lines whereas the binding curve
of Fig, 4B shows monotonic curvature. ~-2-Aminopropanol was employed in preference to 2-aminoethanol, as the kinetic parameters for the reaction with
the latter substrate make it difficult to make precise
measurements with the apparatus at our disposal.
Thus, the values of k + 2 and k t 3 (Scheme 2) are 336 sC
and 240 s C 1respectively with 2-aminoethanol as substrate 181, so the half-life of the decay of the cob(1I)alamin intermediate in the single-turnover experiments
with [El0 [S]o would be abaut 3 ms, which is of
similar magnitude to the dead time of the stoppedflow apparatus.
In the present study the same value of six for the
number of sites per enzyme molecule was found by
using two independent methods : the amplitude and
rate constant methods respectively (see above and
Appendix). This finding is in contrast to that reached
by Babior and by Babior and Li [3-51 who found
that there were two binding sites for coenzyme per
molecule of ethanolamine ammonia-lyase on the basis
of titration studies involving measurement of C D
spectra as a function of AdoCbl concentration, enzyme inhibition as a function of hydroxocobalamin
concentration, binding of 2-aminoethanol in the
presence of methylcobalamin and kinetic measurements with 2-aminoethanol as substrate.
It is difficult to reconcile completely these two
values for n. Part of the difference arises from a difference in the methods for determination of protein
concentration between that of Babior [3- 51 and that

of the present communication. We have employed

dry weight measurements as the basis of our estimations as they are not subject to problems arising from
assumptions made in determinations using the Lowry
or biuret procedures [l 1,121 where protein concentrations are related to an arbitrary standard such as
bovine serum albumin. During the course of these
studies it was found that, by employing the same
procedure as that used by Babior [3- 51, the estimate
of protein concentration would require multiplication
by a factor of about 0.58 to coincide with the value
obtained using dry weight measurements. If this correction factor is applied to the data of Babior his
results give a value of about 3.5 sites per molecule
rather than 2.
The remaining difference of about 2.5 sites per
molecule could arise from a number of sources, including differences in enzyme homogeneity ; different
amounts of residual bound cobamide frequently obtained in preparations of the enzyme); failure of all
the active sites to function with 2-aminoethanol as
substrate (e.g. half-of-the-sites reactivity [17]), or negative cooperativity of coenzyme binding. In the present
study care was taken to ensure essentially complete
saturation of sites by employing AdoCbl concentrations of 63 yM, well in excess of the dissociation
constant for coenzyme of 0.4 pM (see Fig. 4B).
However, a finding that supports the value of
n = 6 is that the enzyme molecule appears to have a
subunit structure that can be designated I&, i.e. the
molecule of M , 520000 probably comprises six subunits of M , 51 000 and six of 36000 [6].
We thank the Science Research Council for financial support
and Mr K . Baker and the Agricultural Research Council Unit for
Nitrogen Fixation for help and facilities for large-scale growth of
clostridia.

REFERENCES
1. Babior, B. M. (1975) in Cobulumiri: Biochemistry and Puthohioiogy (Babior, B. M., ed.) pp. 141-212, John Wiley and
Sons, New York.
2. Kaplan, B. H. & Stadtman, E. R. (1968) J . Biol. Chem. 243.
1793- 1803.
3. Babior, B. M. & Li, T. K. (1969) Biochezistry, 8, 154- 160.
4. Babior, B. M. (1969) J . B i d . Chem. 244, 2917-2926.
5. Babior, B. M. (1969) J . Bid. Chem. 244, 2927-2934.
6. Wallis, 0. C., Johnson, A. W. & Lappert, M. F. (1979) FEBS
L , ~ t t 97,
.
196-199.
7. Joblin, K. N., Johnson, A. W., Lappert, M. F., Hollaway, M.
R. & White, H. A. (1975) FEBS Lett. 53, 193-198.
8 . Hollaway, M. R., White, H. A,, Joblin, K . N., Johnson, A. W.,
Lappert, M. F. & Wallis, 0. C.(1978) Eur. J . Biochem. 82,
143- 154.
9. Kaplan, B. H. & Stadtnian. E. R. (1968) J. Biol. Chern. 243,
1787- 1793.
10. Joblin, K . N., Johnson, A . W., Lappert, M. F. &Wallis, 0. C.
(1976) Biochim. Biophys. Acta, 452, 262- 270.
11. Gornall, A. G., Bardawill, C. J. & David, M. M. (1949) J .
B i d . Chem. 177, 751-766.

M . R. Hollaway, A . W. Johnson, M. F. Lappert, and 0. C. Wallis

185

12. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J.

15. Hull, W. E., Sykes, B. D. & Babior, B. M. (1973)J. Org. C1rern.

38, 2931 - 2939.
16. Babior, B. M., Carty, T. J. & Abelcs, R. H. (1974) J . Biol.
Chem. 249, 1689- 1695.
17. Seydoux, F., Malhotra, 0. P. & Bernhard, S. A. (1974) Crir.
Rev. Biochem. 2, 221-257.

(1951)J. B i d Cirem. 193,265-275.

13. Chance, E. M., Curtis, A. R., Jones,I. P. & Kirby,C. R. (1977)
F A C S I M I L E , Publication of UKAEA, Harwell, No. AERER8775, H. M . Stationery Office. London.
14. Kolthoff, I. M., Shore, W. S., Tan, B. H. & Matsuoka, M.
(1965) Anal. Bioclzem. 12, 497-508.

M. K. ITollaway, Department of Biochemistry, University College London,

Gower Street. London, Great Britain, WCIE 6BT
A . W. Johnson, M. F. Lappert, and 0. C. Wallis, School of Molecular Sciences, University of Sussex,
Falmer, Brighton, Sussex, Great Britain. BN1 9 0 5

APPENDIX

A Kinetic Metliod,for Determining the Number

of Active Sites in an Enzymic Moltwd<~

system and [S]". Second, with [S]" B M . [El0 and fixed

[El()then as [SIC,-+ 'y3 the extent of formation of intermediates will be a function of enzyme concentration,
of n and of the values of the first-order rate constants.

Michael R . Hollaway
This appendix is concerned with a description of
how essentially kinetic methods can be employed to
determine the number of functionally equivalent active
sites, n, in an enzyme molecule. Reference is made in
particular to an enzymic reaction that can be described by Scheme 1 as this is the kinetic model that
applies to the system in the main paper.
EC+S

"=

ECS

1%

ERP G E C + P

(Scheme 1).

In Scheme 1, an active site, EC, in an enzymic molecule containing n such equivalent and independent
sites, combines with a substrate, S, in a pseudoequilibrium process to give a complex, ECS, which
undergoes a reversible transformation to an intermediate ERP, which in turn breaks down to give the
product, P and the unmodified site.
The procedure for the determination of the value
of n is based on analysis of two sets of kinetic data:
one set comprising time courses with the concentration
of active sites much greater than that of substrate
(n . [El0 % [SIC,)and the other with [SI0 9n . [E]o.
The kinetic analysis is concerned mainly with cases
where there is an observable signal change on going
from one enzymic species to another and in particular
for a change in absorbance for the formation of the
cob(I1)alamin intermediate (ERP) from the enzymeAdoCbl-substrate complex (ECS). However the approach can be extended to apply to the case where
signal changes associated with the formation of product are followed.
The principles of the method are based on the following considerations. First, with n . [ElO% [S]Oand
a fixed [S]0, then as [El0 3 GO the extent of formation
of enzymic intermediates and the rates of their interconverions will become independent of enzyme concentration and the value of n, being dependent only
on the values of the first-order rate constants in the

Solution o j the Dffcwntial Equations Arising

.fi.orn Scheme 1

Under this condition the differential equation for

the concentration of the cob(I1)alamin intermediate,
[ERP] derived from Scheme 1 is :
d2 [ERP]
+a.
dt2
where: a

'ERP1
dt

( k i 2+ k-2

+ h . [ERP] = 0

(1)

+ k+3)

and

k i 2 . k+3.

All sites are assumed to function at equal rates. In

addition it is assumed that the k - 3 step can be
neglected. This is a justifiable assumption in the reaction catalyzed by ethanolamine ammonia-lyase (see the
main paper) as the product, P, corresponds to an aldehyde-ammonia addition compound, which will undergo a very rapid breakdown to the aldehyde and
ammonia under the conditions of the experiments [l].
In other systems it might not be possible to assume
such an overall effective irreversibility. However, in
such cases the principle of the method would still be
applicable but the analysis would be complicated by
the presence of terms in k - 3 . [PI.
Use of Laplace-Carson transforms [2] gives the
following transformed equation :

Number o f Active Sites in Lthanolamine Ammonia-lyase

186

where Q is employed in place of the more usual P

to represent the transform operator and
CI

c1 =

(a2

4h)2
; c2
2

~~

(a2

- 4h)2

There are two solutions arising from Eqn ( 2 ) ,one

for r l f r2 and the other for c1 = c2 and these are as
follows.
For unequal c I and c 2 :

where [ERP], is the concentration of the cob(I1)alamin

intermediate at any time t. If k - 2 = 0 then Eqn (3)
simplifies to

corresponding to Eqn (1) of the main paper. Note

that Eqns (3) and (4) are identical in form so the
presence or absence of a 1 k 2 step could not be inferred
from the kinetics under these conditions.
For equal c1 and c2, the appropriate transform of
Eqn (2) gives:

[ERP], = k ; .~[S]O.f . e d r

(5)

with c1 = c2 = c.
For the latter condition to hold the term (a2 - 4h)2
must be identically equal to zero, i.e. the condition
must be satisfied that:

(kL2)

+ 2(k-z - k+3). k;z +

(k-2

+ k+3) = 0

(k-2 - k+3) i (- 4k-2k+3).

(7)

This condition can only hold if k-2 = 0, otherwise

k;2 would have an imaginary component. Therefore
Eqn (7) reduces to kL2 = k+3 and Eqn (5) becomes:
[ERP], = k i 2 . [S]O. t . e-?

(8)

h ) With ( S J o9 n . ( E l o
In this case the formation of the cob(I1)alamin
intermediate will be a pseudo-first-order process if it
is assumed that the concentration of substrate is invariant during the time of observation. Under this
condition integration of the appropriate first-order
differential equation gives :

where

+ k-2 + k + 3 )and kY2 = k+z/(l + Ks/[Sl0).

g = (kY2

Inspection of Eqns (3 - 5 ) shows that, in the experiments with n . [El0 [S]o the Concentration of the
cob(1l)alamin intermediate will increase to a maximum value and then decline. Differentiation of Eqn (3)
leads to the condition for the maximum that it will
occur at time T when:

at which time the maximum concentration of cob(I1)alamin intermediate is given by:

When k-2 = 0 this simplifies further, although is

should be noted that the equation is still of the same
form :

and gives the maximum amplitude arising from Eqn(4).

The equal roots case, which arises only when
k;2 = k+3 and k - 2 = 0, leads to the elegantly simple
solution for a maximum at time T, where
T = l / k + 3 = l/k;2

(13)

in which case:

(6)

from which quadratic we can write:

k;2

The Amplitude Method for the Determination of n

For the experiments with [S]O9 n [El0 Eqn (9) is

applicable, from which equation it can be seen that
the concentration of the cob(I1)alamin intermediate
will approach a maximum as the exponential term
vanishes with increasing time. The expression for this
maximum, steady state concentration, [ERP],,, is :

k;12 . n . [E]p
; -[ERP],, = (kt2
k-2
k+3)

In practice the experimental observations are of

absorbance changes or some other signal rather than
of concentrations so the molar absorption cofficients
have to be introduced and these are, in general, unknown quantities. However, if it is assumed that the
only spectral change occurs on going from the enzymeAdoCbl-substance complex to the cob(1I)alamin intermediate (for example), then the molar difference absorption coefficient, A s , for this change can be eliminated by using the relationships : A,,, = d E [ERP],,,
and A,, = d E [ERP],, in combination with Eqns (1 1)
and (25) to give Eqn (16):

M. R. Hollaway, A. W. Johnson, M. F. Lappert, and 0. C. Wallis

187

where the subscript to the concentrations (1 and 2)

refer respectively to the 'transient' and 'steady-state'
experiments and :

Determination of n
from Relationships Involving Observed Rate Constants
a ) With k - 2

In the case of equal roots, use of Eqns (14) and (15)

gives :

In the derivation of Eqn (18) ky2 has been set equal

to k+2 (i.e. the assumption has been made that
[SIO 9 Ks).
The following simplifications arise from Eqns( 16 18) when the conditions are imposed that: [E]o,l 9 K,
and [S]O,Z9 K, respectively, because in that case
ki;2 = k>2 = k+2.
a) If k-2 = 0, (the assumption made in the main
paper), then the expression for , f ( k )simplifies greatly
such that :

Inspection of Eqn (19) reveals that if either k+2 3 k+3

or k+2 < k+3 then f ( k ) z 1. That is, under these
circumstances Eqn (16) simplifies to:

and so a value for n can be calculated from the four

measurable or known entities without any further
knowledge of the magnitude of the rate constants.
In between the extremes in the relationship between
the values of k+2 and k+3 it can be shown that the
smallest possible value of f ( k ) of 2/e (i.e. 0.736) is
given by the 'equal roots' case [see Eqn (18)]. Thus,
the 'complicating factor', f ( k ) , that enters into the
calculations lies between 1 and 0.73 so the amplitude
treatment would give a 'rough' estimate for the value
of n without the knowledge of the detailed kinetics
of the system.
b) If k - 2 is finite but the conditions that k+z 9 k+3
and k+2 9 k-2 apply thenJ'(k) will still be close to
unity because in that case c2 % c1 [see Eqn (17)]. In
the experiments described in the main paper the spectrum observed during the steady state is close to that
expected for a cob(I1)alamin intermediate, with little
or no Co(II1) species present, so k-2, if finite, must
be much lower in value than k+2. This requires also
that k+2 9 k + 3 .Thus Eqn (20) could be applied with
little error although, as the values of the rate constants
were known, the more refined treatment based on
Eqn (19) was carried out.

=0

If k+2 9 k+3 also, then Eqn (4) applies, with the

rate constant for ERP formation in the experiment
with n . [El0 9 [S]o corresponding to k + 2 and that for
its breakdown to k+3. In the experiment under the
steady-state conditions the maximum velocity will be
given by: V = n . k + 3 . [ E ] O , 2 , so enabling a value for
n to be calculated by substitution of the measured
values for V, k+3 and [ E ] o , ~Conversely,
.
when k+2
4 k+3, a consideration of Eqn (4) reveals that in the
transient type of experiment, the cob(I1)alamin intermediate will also build up and decay but at the
maximum a very small fraction of the substrate will
be in the form of intermediate and the apparent rate
constant for formation of the intermediate will be
k + 3 (i.e. correspond to the rate constant for the
breakdown) and that for its breakdown will be k+2
(i.e. correspond to the formation constant). If the
sensitivity of the apparatus allows the transient formation of a small amount of the cob(1I)alamin intermediate to be followed, a value for n could then be
calculated by substitution of the value for k + 2 obtained from the breakdown constant, in the relationship for maximum velocity, V = n . k+2 . tE10.2.
The shape of the curve arising from the 'equal
roots' case [Eqn (S)] would present practical problems
in the treatment of the kinetic data as a first-order
rate constant cannot be determined from either the
'formation' or 'breakdown' part of the curve (e.g.
10 half-lives must elapse before the experimenally observed rate constant for the degradation would be
within 10% of the correct value).

h ) With k - 2

+0

If k+2 % k+3 the system is still amenable to an

analytical treatment. With n . [ElO9 [Sl0 Eqns (2)
and (3) will apply but with c2 CI. Thus the exponential for the formation of the cob(I1)alamin intermediate, e-"' , will vanish before the breakdown
exponential, e-"' , changes significantly and so the
rate constant for the decay of the intermediate will
correspond to c l . The expression for the latter is :

c1 =

![l
2

(1

where u = (k+2+ k-2 + k+3) and h = k+2 . k+3.

AS k + z % k+3, then a' 9 4h, allowing a binomial
expansion of 1 - 4h/a2 and the neglect of terms containing squared or higher powers of 4 h / a 2 ,whence:
CI %

k+2 . k+3/(k+2 k-2

+ k+3)

kYpp.

(22)

188

M. R. Hollaway et al.: Number of Active Sites in Ethanolamine Ammonia-lyase

The steady-state expression for the maximum velocitv is:

An appropriate substitution in Eqn (23) allows the

calculation of a value for n.
In principle the kinetic methods for determination
of the number of functional active sites in an enzyme
molecule could be extended to apply to any system
where it is possible to follow the time course of formation and breakdown of an enzymic intermediate.
Clearly, in cases where only the decay of substrate

or formation of product can be monitored, oiily the

procedure involving measurcinent of rate constant$ is
applicable. In any case the analysis presented here
shows that it is necessary to have an understanding
of the kinetic mechanism of the reaction and a knowledge of at least the relationships between rate constants before a reliable measurement for I I can be
ensured.

REFERENCES
1. Hull, W. E., Sykes, B. D. & Babior, B. M. (1973) J . Org. Chem.
38, 2931 -2939.
2. Rodiguin, N. M. & Rodiguina, E. N . (1964) in Consecutive
Chemicul Reactions (Schneider, R . F., ed.) pp. 109-134, D.
van Nostrand Co. Inc., Princeton, New Jersey.

M. R. Hollaway, Department of Biochemistry, University Collcgc London,

Gower Street, London, Great Britain, WClE 6BT