BIOTECHNOLOGY AND BIOENGINEERING

VOL. XVI, PAGES 904-923 (1974)

External and Internal Diffusion in Heterogeneous

Enzymes Systems
CSABA HORVATH and JEAN-MARC ENGASSER, Department
of Engineering and Applied Science, Section of Physical Sciences,
School of Medicine, Yale University, New Haven, Connecticut 06620

Summary
The intrusion of diffusion in heterogeneous enzyme reactions, which follow
Michaelis-Menten kinetics, is quantitatively Characterized by dimensionless
parameters that are independent of the substrate concentration. The effects
of these parameters on the overall rate of reaction is illustrated on plots commonly
employed in enzyme kinetics. The departure from Michaelis-Menten kinetics
due t o diffusion limitations can be best aasessed by using Hofstee plots which are
also suitable to distinguish between internal and external transport effects.
A graphical method is described for the evaluation of the reaction rate BS a
function of t,he surface concentration of the substrate from measured data.

INTRODUCTION
The transport of substrate to the catalytic sites has often been
found to affect the overall kinetic behavior of immobilized or membrane-bound enzymes. The effect of both external and internal substrate diffusion on the Michaelis-Menten kinetics has been theoretically treated by using different appro ache^.'-^ The purpose of this
study is to treat the intrusion of diffusional effects by dimensionless
parameters which are independent of the substrate concentration.
Such parameters make it convenient to illustrate the effect of diffusional limitations on graphs commonly used in enzyme kinetics.

EFFECT OF EXTERNAL TRANSPORT

When an enzymic reaction takes place a t a surface, substrate
transport occurs by molecular or convective diffusion. In the following treatment it is assumed that the enzymic reaction per se
obeys a Michaelis-Menten kinetic law and the catalytic surface is
equiaccessible, i.e., the conditions for substrate transport to any
909

@ 1974 by John Wiley & Sons, Inc.

910

HORVATH AND ENGASSER

point on the surface are identical. At steady state the rate of substrate transport to the surface is equal to the rate of substrate consumption by the reaction, i.e.,

where C , and C b are the surface and bulk concentrations of substrate,

respectively, h. the transport coefficient, V,,," the saturation rate per
unit surface area, and K mthe Michaelis-Menten constant.
By introducing the dimensionless concentration, j3, defined as

C/Km

(2)

and a dimensionless group which is often referred to as a Damkoehler

number, Da, given by
Da

Vmax"
hK,

= __

(3)

eq. (1) can be written in dimensionless form as

where @ b and
are the dimensionless bulk and surface concentrations, respectively.
Equation (4)shows that two limiting cases are possible. When,
at a given O b , the Damkoehler number is sufficiently large,
goes
to zero because the reactivity at the surface is so high that all substrate molecules which can be transported to the surface are immediately converted. As a result the observed reaction rate is equal to
the maximum possible rate of external transport, V d i t f n , at a given
bulk concentration, i.e.,

v"

= vdiff''

hCb

(5)

In this case the measured reaction rate is virtually independent of

the kinetic parameters of the enzymatic reaction.
When Da is very small, the bulk concentration is practically equal
t o the surface concentration. Then, the observed reaction rate is
equal to the maximum possible rate of the actual chemical reaction,
V~i,,", a t a given bulk concentration, thqt is,

DIFFUSION IN ENZYME SYSTEMS

911

The two limiting rates, v d i f f n and V k i n , therefore, can be considered

as expressions for the conductances of the transport process and the
reaction, respectively.
When the surface is uniformly accessible and the overall reaction
rate depends on both external diffusion and chemical reaction, the
two processes take place in series as if they were independent of each
other. At steady state both must have the same rate that is determined by both conductances, V d i f f N and V k i n . If, however, one of
the conductances is much smaller than the other for a given concentration, then the overall rate, V, which would be measured experiment,ally,is limited by the process of lower cond~ctance.~In Figure
1 the situation is illustrated by the dependence of V k i n , V d i f f , and
the resulting V on the bulk concentration, for a set of arbitrarily
chosen kinetic and diffusional parameters. At sufficiently high concentrations, V,,,, the saturation value of V k i n , can always be obtained. At lower concentrations either V d i f f or V k i n plays the predominant role in determining the overall rate of reaction. The relative effect of these limiting rates depends on the value of Da, i.e.,
on the relative magnitudes of the initial slopes of V d i f f n and V k i n .
The limiting rate, V k i n , is obtained over the whole concentration
range if external diffusion is fast enough to maintain a flat concen-

BULK CONCENTRATION

Fig. 1. Plot of the overall rate of the surface reaction, V, against the substrate
concentration in the bulk solution. In this case, V is determined by both
the limiting rate of the actual enzymic reaction at the surface, Vkin, and the
maximum possible rate of substrate diffusion to the surface, Vdiff, which are
also illustrated.

912

HORVATH AND ENGASSER

1.0

0.8

0.2

1000
0.0

10

20

30

40

50

4
Fig. 2. Plots of the overall rate of an enzyme catalyzed surface reaction, V,
normalized to V,.,
against the dimensionless bulk concentration, 06,at different
values of Damkoehler number, Da. The effect of diffusion limitations increases
with increasing values of Da and results in a decrease of the overall rate of
reaction with respect to the kinetically controlled rate, Vkim/Vmx, which
is obtained at Da = 0.

tration profile. When the transport of substrate is relatively slow,

V departs from VkinNas shown in Figure 2 where the normalized overall reaction rate, V/V,,xfl is plotted against pa at different values
of Da. As seen, at large Da values the overall rate of reaction is
lower than that of the intrinsic enzymic reaction shown for Da -+ 0,
because the surface concentration of the substrate is smaller than the
bulk concentration due to external diffusion limitations. Such effects
are negligible only if Da is smaller than unity. Nevertheless, at
sufficiently high concentrations the depletion of the substrate at the
surface region does not affect the overall rate significantly, even at
large values of Da, so that the saturation rate, V,,,, can always be
attained, at least theoretically.
In order to illustrate the effect of kinetic limitations on the overall
reaction rate due to relatively low catalytic activity, V is normalized
to the product hK,, which is proportional to Vdiff. In Figure 3 this
normalized rate is plotted against @ b with Da as the parameter. It
is seen that at any given Da value, the departure of V from Vdjff
increases with the bulk concentration because the rate of transport

DIFFUSION IN ENZYME SYSTEMS

913

50

40

30
E

C
:
\

>

20

10

0
0

10

20

30

40

50

4
Fig. 3. Plots of the overall rate of an enzyme catalyzed surface reaction,
normalized to the product of the mass transfer coefficient,h, and K,, against the
dimensionless bulk concentration, O b , a t different values of the Damkoehler
number, Da. The effect of kinetic limitations increases with decreasing values
of Da and results in a decrease of the overall rate of reaction with respect to the
diffusion controlled rate, VdiffnlhK,,,, which is obtained at Da = 00.

increases to a greater extent than the rate of the enzymic reaction

and only a fraction of the substrate molecules that could diffuse to
the surface are converted. On the other hand, a t any given bb
value, V is approximately equal to Vdiff only if Da is sufficiently
large, i.e., the enzyme activity is high enough to maintain the surface
concentration practically zero. Figure 3 shows that at f i b < 5 the
10, but a t higher
reaction is already diffusion controlled when Da
bulk concentrations V d i f r N is reached only at much higher values
of Da.
The interaction between external diffusion and surface reaction
can be expressed by the departure of V from Vkin or VdiffN in view
of the foregoing discussion. For convenience an effectiveness factor,
q e is defined by

>

V = qeVkin

(7)

which is a quantitative measure of the effect of external diffusion on

the overall reaction. Clearly, q s is unity for a kinetically controlled
reaction and its value decreases with increasing diffusion limitations.
The dependence of q e on p b and Da is shown in Figure 4. When

HORVATH AND ENGASSER

914
I

.I

01

10

100

1000

Da

Fig. 4. Graph illustrating the external effectiveness factor, qe, as a function

of the Damkoehler number, Da, with the dimensionless bulk concentration, f i b ,
aa the parameter, for an enzymic surface reaction. At sufficiently small concentrations the limiting first order effectiveness factor ee, is reached.

< 0.1, then q e becomes independent of the concentration, as

expected for a first order reaction, and approaches a limiting value,
e e , which is related to Da by

@b

e, =

+ Da

In Figure 4 the locus of the kinetically controlled reaction is at

1, on the upper horizontal line, whereas the straight lines obtained for q e at high Da values represent the diffusion controlled
reaction domain. The border line between the diffusion controlled
and the intermediate domain of the effectiveness factor, which is
affected by both diffusion and kinetics, is indicated by the broken
line. Thus there is a critical Da for any bulk concentration where
t,he reaction becomes diffusion controlled for all practical purposes,
so that any increase in Da has no significant effect on the overall
rate. Since, at a fixed h, the increase in Da is equivalent to an increase in the amount of enzyme, in practice the maximum possible
rate can be obtained with some minimum amount of enzyme att,ached to the surface.6 The corresponding optimal ,T,I
value can

qc =

DIFFUSION IN ENZYME SYSTEMS

915

be calculated for a given f i b , h, and K m by evaluating Da a t the

broken line.
Although the dependence of V on the bulk concentration is usually very complex as illustrated in Figures 2 and 3, it can be easily
treated in the following two limiting cases. At a sufficiently high
bulk concentration both V and Vkinb have the same limiting value,
V,,,, the saturation rate of the enzymic reaction. On the other
hand, a t low concentrations V is a first order reaction with a rate
constant V m a x N / ~where
e
K~ is defined by

It should be noted that eqs. (3), (8), and (9) can be combined to
obtain the following equation:

which expresses the resistance additivity relation for the first order
enzyme reaction cum external transport.
This similarity in the limiting behavior of V and v k i n , however,
does not necessarily mean that the functional relationship between
1 and the concentration also has the form of a rectangular hyperbola that is characteristic of the Michaelis-Menten kinetic law. This
is illustrated in Figure 5, which shows plots of the normalized overall
reaction rate, V/Vmnx,against V/(Vm,, x f i b ) . Such plots, like
the so-called Hofstee plots used to evaluate kinetic data obtained
with soluble enzymes,* yield straight lines when V obeys the
Michaelis-Menten law, tts can be seen for Da + 0 in Figure 5 . However, as a result of diffusion limitations, curves corresponding to increasing values of Da, depart significantly from straight lines, particularly when a wide concentration range is examined. These
portions of the plots where the overall reaction rate is essentially
determined by the rate of external transport are straight vertical
lines.
The assumption that external transport and surface reaction take
place in series can be used to obtain thetrue kinetic rate Vkin(C8)
from the measured rate V ( C b ) . When the mass transfer coefficient
for the substrate, h, is known, the concentration of the substrate at,
the surface can be determined from the bulk concentration and the
reaction rate by using the following relationship :

V(Cb)

vkin(C8)

h(cb - Cs)

(11)

HORVATW AND ENGASSER

916
1.0

0.25

0.5

0.75

0.75

0
x

=E

>
>

=\

0.5

0.25

0.0
0.0

1.0

VI;IV;axXpb
Fig. 5. Departure from Michaelis-Menten kinetics due to external transport
limitations as illustrated by Hofstee type plots at different Damkoehler numbers,
Da. The overall surface reaction rate, V, obtained at different dimensionless
bulk concentrations, P b , is normalized to V-,. In the diffusion controlled reaction domain, i.e., at large values of Da the plot yields vertical straight lines.

A graphical method based on eq. (11) for the determination of Vkin

is illustrated in Figure 6. First, the measured V is plotted against
the bulk concentration, c b . Then, a t any chosen Ca a straight line
with the slope equal to - h is drawn, which represents the flux of the

substrate to the surface for all c, values from zero to the given c b .
This straight line intersects the horizontal line, which is drawn to
represent the value of V at this particular C b . The intersection
yields both the actual surface concentration, C,, and. the intrinsic
rate of reaction, V k i n , for that value of c b . Thus, a plot of V k i n
against the surface concentration can be constructed by repeating
the procedure for a number of bulk concentrations. From this plot
the kinetic parameters of V k i n can then be determined by standard
methods. This graphical procedure is not restricted to MichaelisMenten kinetics and is quite generally applicable provided the assumption of equiaccessible surface is appropriate and the external
transport coefficient is known.

DIFFUSION I N ENZYME SYSTEMS

917

In the caae of a porous enzymic medium with internal diffusion

limitations this method yields the overall rate as a function of the
surface concentration, which is considered in the following section.

EFFECT OF INTERNAL DIFFUSION

When enzymes are embedded in a porous medium, the substrate
has to diffuse inside the pores in order to react a t the catalytic sites.
Unlike external transport, internal diffusion proceeds in parallel with
the chemical reaction. Thus, the concept of limiting processes in
series is no longer applicable and the treatment of the subject is more
complex. Our model consists of a membrane containing bound enzymes uniformly distributed. The membrane can be exposed to the
substrate solution on one side and sealed on the other. Alternatively,
both sides can be open and in contact with the same substrate concentration. At steady state the concentration change in the membrane
is described by the following differential equation :

where Deffis the effective diffusivity of the substrate in the memis the maximum
brane, C is the local substrate concentration, Vmadtt
possible reaction rate per unit volume of the membrane, and x is the
distance from the surface. By defining the dimensionless position
in the membrane, z, as
x
z = -

(13)

where 1 is the thickness of the membrane, eq. (12) can be written in

dimensionless form as

d28 =

92

dz2

B
~

1+B

In eq. (14) 4 is the so-called Thiele modulus given by

Numerical integration of eq. (14) with the boundary conditions

B=Bd

atz=O

(16)

and
clSJdz = 0

at z = 1

(17)

918

HORVATH AND ENGASSER

CONCENTRATION

Fig. 6. Graphical determination of the surface concentration, C,, and the

rate of the actual kinetically controlled enzymic reaction a t the surface, Vkin,
from the. overall rate of reaction V, measured at various substrate concentrations in the solution, Cb, when the mass transfer coefficient, h, is known. At a
chosen value of c b a straight line of slope -h is drawn. The intersection of this
line with a horizontal line drawn at the value of V measured at the same c b
yields both the corresponding surface concentration on the abscissa. and the
intrinsic rate of the reaction for this surface concentration on the ordinate. By
repeating this procedure for different Cb values a plot of Vkin against C. can be
constructed.

yields the substrate concentration profile in the enzymic membrane

and, consequently, the overall rate, V , for the whole enzymic membrane can be calculated. This rate normalized to the maximum
possible rate, V,,,, is plotted as a function of the surface concentration with the Thiele modulus as the parameter in Figure 7. It is
seen that for 4 < 1 the reaction is essentially kinetically controlled,
but a t higher values of 4 the observed rate is lower due t o diffusion
limitations unless the saturation rate, V,,,, is reached at high concentrations. As expected V is always a function of the kinetic
parameters of the reaction even at very high values of cp, in contrast
to the case of external diffusion, when the overall rate is independent
of the kinetic parameters a t sufficiently high D a values.
The dependence of V on the surface concentration is usually complex as seen in Figure 7. At small surface concentrations,
< 0.1,
however, the overall reaction rate is first order with the rate constant
V r n a x / ~where
i
~i is defined by
Kmcp
Ki =

tanhcp

DIFFUSION IN ENZYME SYSTEMS

919

I.o

0.0

0.6

E
>

0.4

0.2

0.0
0

10

20

30

40

50

Ps
Fig. 7. Plots of the overall rate of reaction in an enzymic membrane, V ,
normalized to V,,,,,, against the dimensionless concentration of the substrate at
the surface, pa, at different values of the ThieIe modulus, Q. The effect of
diffusion limitations increases with increasing values of Q and results in a decrease
of the overall rate of reaction with respect to the kinetically controlled rate
which is obtained at Q = 0.

On the other hand at sufficiently high values of B8, internal diffusion

has no effect on the overall rate since V becomes equal to the saturation rate, Vmax. Thus the observed rate shows a limiting behavior
similar t o that of the Michaelis-Menten scheme. Yet, its functional
dependence on the surface concentration is different as illustrated
by the plots in Figure 8, which deviate from straight lines when
4 > 1.
The effect of internal diffusion on the kinetic behavior of immobilized enzymes in spherical particles has been demonstrated by plotting the calculated overall reaction rates according t o the schemes
known as Lineweaver-Burk and Hofstee plots4. The deviations from
straight lines due to the departure from the Michaelis-Menten
kinetics have been most pronounced in the latter type of plots, even
at relatively low values of the Thiele modulus, in agreement with the
earlier suggestiong that the Hofstee plot is more suitable t o discern
deviations from the straightforward Michaelis-Menten kinetics.
In this study this plot has been found also very sensitive to the
effect of internal or external diffusion limitations which manifest

920

HORVATH AND ENGASSER

1.0

0.75

E
>
>

0.5

0.25

0.0
0.0

0.25

0.5

0.75

1.0

~ ~ ~ I n4 , , s

Fig. 8. Departure from Michaelis-Menten kinetics in an enzymic membrane

due to diffusion limitations aa illustrated by Hofstee type plots obtained at
different Thiele moduli, 9. The overall reaction rate, V , obtained at different
surface concentrations, &, is normalized to Vmax. Diffusion limitations at
nonzero values of 6 result in sigmoid curves.

themselves by vertical lines and sigmoid curves, respectively, as

shown in Figures 5 and 8. Although Lineweaver-Burk plots are
used most commonly in the study of immobilized enzymes, the
results strongly suggest that Hofstee plots are most appropriate for
the investigation of the kinetic behavior of heterogeneous enzyme
systems. Such plots are not only more sensitive t o the effect of
diffusion limitations but also yield information about the nature of
the diffusional effect by analyzing the shape of the curves obtained
with experimental data. Furthermore, Hofstee plots can also supply
valuable information on the combined effect of product inhibition
and transport phenomena on the kinetics of heterogeneous enzyme
systems.10
The results of this study clearly indicate that the Michaelis-Menten
kinetic law is no longer obeyed when transport limitations affect the
enzymic reaction ; consequently, there is no K , value to characterize
the overall kinetics of the reaction under such conditions. In view

DIFFUSION IN ENZYME SYSTEMS

921

of this fact it seems t o be appropriate to review the different ways

proposed in the literature to quantitatively express the effect of
transport limitations in heterogeneous enzyme kinetics.
When transport phenomena affect the enzymic reaction i t is generally recognized that the overall rate no longer follows the MichaelisMenten kinetic law. Nonetheless, apparent K, values have been
widely used and suggested to characterize the intrusion of diffusional
effects in the kinetics of the reaction. Furthermore, such apparent
K , values have been defined in different ways. I n some cases apparent K , values were used to express the limiting first order constants,11J2 which are equivalent to K , and K~ discussed above. I n
other cases the apparent K , was defined as the concentration a t
which the rate of the heterogeneous enzyme reaction is half of the
saturation rate.13J4
Although both apparent K , values are the same and equal t o the
actual K , in the absence of diffusion limitations, they are significantly
different a t larger values of 4 and D a as shown in Figures 2 and 7
as well as by eqs. (9) and (18). I n view of this, the use of the term
apparent K , value for the parameters denoted by K in this paper or
the substrate concentration yielding the half saturation rate does not
appear to be appropriate when the overall reaction does not follow
the Michaelis-Menten kinetic law. Nevertheless under certain conditions, e.g., due t o some microenvironmental phenomena or the
effect of transport limitations and product inhibition in some particular casesl10 the overall kinetics may formally follow this law
despite the intrusion of physical and chemical phenomena. Even
then the use of a n experimentally obtained apparent K , value can
be quite dangerous in that i t inspires false confidence and could conceivably be used in design under circumstances when the fluid mechanical conditions are far removed from those employed in the kinetic experiment.

APPENDIX
The reviewer pointed out that there is no optimum amount of immobilized enzyme in the external diffusion controlled system as
claimed in this paper because the overall reaction rate always increases by increasing V,,,, i.e., the amount of the surface-bound
enzyme. The proof provided is as follows.
If only VmaXis varied i t can be shown t.hat

922

HORVATH AND ENGASSER

Since the right-hand side of eq. (19) is always positive, bV/bV,,,

is always greater than zero, i.e., the rate always increases with V,,,.
In the rigorous theoretical sense that is true because VdirfN as well
as v k i n are fictitious reaction rates. In theory Vdirr would only be
reached at Da = CQ to obtain zero surface concentration. Nevertheless Figure 3 shows that in practice Vdirfcan be obtained in a wide
bulk concentration range already at Da = 100. Then C , and,
according to eq. (19), bV/bV,,, are practically zero and a further
increase in Da by increasing V,,, does not yield a higher reaction
rate.
Similarly, Vkin is never reached theoretically because Da and the
effectiveness factor is never zero and unity, respectively. According to our analysis, however, the reaction rate equals Vkinin practice
when Da < 0.01.
This conflict of asymptotic limits, which are never obtained theoretically and limiting rates which are encountered in practice, is
inherent to the Michaelis-Menten kinetic scheme also. For that
case, if only the concentration is varied, it can be shown that

As seen, bV/bC is always

greater
than zero so that the saturation
.
V,,,, is never reached theoretically.

Nomenclature
substrate concentration
substrate concentration in the bulk fluid
substrate concentration at the interface
substrate diffusivity in external solution
effective substrate diffusivity in the enzymic medium
Damkoehler number
mass transfer coefficient for the substrate
Michaelis-Menten constant
thickness of membrane
overall reaction rate
overall reaction rate per unit surface area
maximum rate of substrate transport per unit surface area at a given bulk
concentration of the substrate
maximum reaction rate per unit surface area at a given surface concentration of the substrate
saturation reaction rate according to Michaelis-Menten kinetics
VI%3X saturation rat,e uer unit area of catalyst surface
V,,, saturation rate per unit volume of catalyst
internal distance from the membrane surface
I
dimensionless internal distance from the membrane stirface
z

DIFFUSION I N ENZYME SYSTEMS

923

Greek Let.ters
j3
j3b

C.

rlC
K~

~i

dimensionless substrate concentration

dimensionless substrate concentration in the bulk fluid
dimensionless substrate concentration at the interface
external effectiveness factor for first order reaction
external effectiveness factor
kinetic parameter for first order enzyme reaction with external transport
limitations
kinetic parameter for first order enzyme reaction with internal diffusion
limitations
Thiele modulus

This work was supported by grants No. R01 GM 21084 and No. RR-356 from
the National Institutes of Health, U.S. Public Health Service. J.-M.E. is a
recipient of a Yale University Fellowship.

References
1. E. Katchalski, I. H. Silman, and R. Goldman, Advan. Enzymol., 34, 445
(1971).
2. M. Moo-Young and T. Kobayashi, Can. J. Chem. Eng., 50, 162 (1972).
3. D. J. Fink, T. Y. Na, and J. S. Schultz, Biotechnol. Bioeng., 15,879 (1973).
4. J. M. Engasser and C. Horvath, J. Theor. Biol., 42, 137 (1973).
5. D. E . Rosner in Annual Review of Materials Science, Vol. 2, R. A. Huggins,
Ed., Annual Reviews, Inc., Palo Alto, Calif., 1972, pp. 573-606.
7. D. A. Frank-Kamenetskii, Diffusion and Heat Transfer in Chemical
Kinetics, Plenum Press, New York, 1969, pp. 53-57.
8. B. H. J. Hofstee, Science, 116, 329 (1952).
9. J. E. Dowd and D. S. Riggs, J. Biol. Chem., 240, 863 (1965).
10. J. M. Engasser, Doctoral thesis, Yale University, 1974.
11. W. E. Hornby, M. D. Lilly, and E. M. Crook, Biochem. J., 107,669 (1968).
12. P. S. Bunting and K. J. Laidler, Biochemistry, 11,4477 (1972).
13. P. V. Sundaram, E. K. Pye, T. M. S. Chang, V. H. Edwards,
A. E. Humphrey, N. 0. Kaplan, E. Katchalsky, Y. Levin, M. D. Lilly, G.
Manecke, K. Mosbach, A. Patchornik, J. Porath, H. H. Weetall, and L. B.
Wingard, Jr., in Enzyme Engineering (BiolechTol. Bioeng. Symp. No. 3) L. B.
Wingard, Jr., Ed., Wiley-Interscience, New York, 1972, pp. 15-18.
14. R . Goldman, 0.Kedem, and E. Katchalski, Biochemistry, 10, 165 (1971).

Accepted for Publication January 14, 1974