Eco. o
.133; 0:1
4

3 2; ‘
Un.3,...
1.4.x» .9
5 P

.
.

r
. II... .. .
.ngP—ai w
v

, . V .. . :$%W§

 

a .x.
r, ¢JsE..
¢

$A

.- ..

_
Wynn; w? r J
hm Emma"

fin...”

. no... . ‘ $3.".
x u§|
3..

.. 4.

.I ..
:u-Tt

5:: a: L’-
|A .

3..
:0:
z

.9!
:1

9. {v £5
. .91.. guiaickfia
I? (a... it. .
11:15) -. ,
a.
a
,9...
u h:
(.13: L
7’!

7Qh1§~‘-"v|l O n W
.2.’ x:a\9.)2
s: L in: 7(. 2
v I 31.5 to? 1:1
1.... 53.. ,.
.....L;Eé.r
3:11;]!

:1“:ka
5. . )Ilh..ir:fl4
.331 ,
(4r: 1!.-
II 3 x , ..l:...!x~i.-:: l}
r l: 3 5.2.3.1.}! a...<{.....i..n
a . xx 5:: :. it 5.5. :
..K. ‘25-; . by

ailfi"!
30:59:.
9r

.SL :5
l. i: .51.!
13.x...

,\ . .X: b 3 .P
If 5'22: 31‘]:
5.35.13. it):
(.2. : ‘a.9'..s..
I}. 5.9

:6.

in,

, ‘ ’n‘ 4:1”.
. : . . . . .. . .5335:

2-..,“ .7. a. ”.10 t..|l.5:!-?

. . ‘ . 3" i-‘fi‘l'

7.. .n‘.

; £3
.;vvn(!l.t:?vu§ll.)as
iraflt‘: . I

:1 1.3.3.3...

G) ;.:..z£!arx\:i . . .. 1..

2?.» ((1.. .25.}..-
.2- .I51

 

£51.?
y . fix}..-
‘1‘:

. .2: . .3. t
2:.) 2 2.2 I W
2". 2.1:; .2. El. 5,3,3:

1‘ .3~L.

,, . , Iz; iéé .. ccfiémm
“3m «{svmwfiafimfimfir .gmfi‘wrm gfigfifiqukwfrz

 

‘uullllgllnle
ZTLO

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

INTERACTIONS 0F SMOOTH MUSCLE AND SKELETAL MUSCLE
PYRUVATE KINASES WITH CREATINE KINASE

presented by

Patrick Robert Sears

has been accepted towards fulfillment
of the requirements for

Ph - D - degree in Phys 10108):

 

 

flflfln

Major professor

Date May 7, 1999

 

MS U is an Aflirmariw Action/Equal Opportunity Institution 0-12771

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11/00 m.“

 

INTERACTIONS OF SMOOTH MUSCLE AND SKELETAL MUSCLE PYRUVATE
KINASES WI I H CREATINE KINASE

By
Patrick Robert Sears

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Physiology

1 999

ABSTRACT

INTERACTIONS OF SMOOTH MUSCLE AND SKELETAL MUSCLE PYRUVATE
KINASES WITH CREATINE KINASE

By

Patrick Robert Sears

Physical and functional coupling of pyruvate kinase (PK) and creatine kinase
(CK) has been previously noted with the skeletal enzymes. An isolation procedure was
developed for PK from smooth muscle in order to investigate the coupling in this tissue.
The coupling of PK with CK in smooth muscle was investigated in relation to the known
coupling in the skeletal muscle.

Kinetic parameters of smooth muscle PK (SMPK) were determined to be similar
to those of skeletal PK (SKPK). The Km's for smooth muscle PK were 214 uM ADP and
86 uM PEP. The Km's for skeletal muscle PK were 254 uM ADP and 40 uM PEP. The
average Vmax for SKPK was 0.068 U/ug and for SMPK was 0.041 U/ug. Arrhenius
plots were linear for both enzymes and the observed activation energy was 61,100 J/mole
for SMPK and 56,400 J/mole for SKPK.

A new method for quantitative measuring of coupling between molecules was
developed using capillary electrophoresis (CE). Using these methods it was possible to
control the degree of coupling independently by either controlling the concentration of the
molecules or the electric field. A CE method was developed for running PK and CK.
Qualitative measurements of PK-CK coupling were achieved using the CE methods.

Procedures which had been previously used for measuring the coupling between
the skeletal enzymes were used'and adjusted to measure the coupling between the smooth

muscle enzymes. These procedures involved the measurement of absorbance changes by

the enzymes and changes in their ethanol solubility. A new effect of CK to increase the
activity of PK was discovered in the course of these experiments.

MMCK is the CK isoform found in skeletal muscle. Smooth muscle has been
shown to contain mostly another isoform, BBCK, and some MMCK. The coupling
experiments showed that SMPK does not couple with BBCK or MMCK, but that SKPK
couples to both. These findings suggest that it may be the isoform of PK that determines
PK-CK coupling in different tissues.

Since glycolytic fluxes in skeletal and smooth muscle are different, the differences
in SMPK and SKPK coupling may reflect features of glycolytic flux control which are
specific to smooth and skeletal muscle. Using the CE coupling methods the electric field
required to uncouple molecules can be determined. The electric field required for
dissociation is smaller than that found at distances (0-10 nm) from membranes which are
several times larger than the diameter of the enzymes. This suggests that the location of

the enzymes at specific sites within the cell may effect their coupling properties.

This dissertation is dedicated to the person who is dearest to me,

Kelly.

Your patience, your caring, your work, and most of all your love;
I will always remember when I think of these times.

iv

ACKNOWLEDGEMENTS

I would like to thank all those who have worked in the lab with me. In my time
here I have only met people willing and happy to help. I would like to thank my
committee members. Doctors Ti Tien, Seth Hootman, and Raoul Hollingsworth for
reviewing my work; and Doctor Robert Root-Bemstein for his advice on all aspects of my
research. I must especially thank Doctor Patrick Dillon for his time helping me in the lab,
his advice in professional and other matters, the hours we spent going over data together,
his support during difficult times, and his incredible patience.

There are many who gave me support outside the lab but without whom I would
not have succeeded. Foremost among those are my parents, Frederick and Nicole Sears--
Thank you for all your care.

TABLE OF CONTENTS

LIST OF FIGURES
LIST OF ABBREVIATIONS

INTRODUCTION

BACKGROUND
The classical view of metabolism
Regulatory enzyme kinetics and similarities with the modern
view concepts
The modern view of metabolism, non-specific interactions
The modern view of metabolism, specific interactions
Glycolysis and the modern view of metabolism

METHODS
Urinary bladder smooth muscle PK isolation
General CE methods used in all sections
Bladder metabolites
Kinetics
CE coupling
PK-CK interactions

RESULTS
Isolation of smooth muscle pyruvate kinase
Kinetic properties of skeletal and smooth muscle pyruvate kinases
CE coupling experiments
PK-CK interactions

vi

Page

viii
xi

14
18
21

25
25
40

50
52

68
68
70
79
87

DISCUSSION
Smooth muscle PK isolation
Properties of SKPK and SMPK compared
CE coupling experiments
Specific properties of pyruvate kinase and odd characteristics
Creatine kinase and channeling
A general view of the cell
The relativity of specific weak interactions
Current protein interactions research

SUMMARY

REFERENCES

vii

93
93
94
95
99
101
104
106
108

110

111

LIST OF FIGURES
Figure 1. Effect of polyethylene glycol concentration on smooth muscle pyruvate kinase
activity.

Figure 2. pH - dependence of smooth muscle pyruvate kinase binding to carboxy-methyl
coated agarose beads.

Figure 3. Non-effects of centrifugation, carboxy-methyl binding, and pH on smooth
muscle PK activity during SMPK isolation.

Figure 4. 280 nm Absorbance and smooth muscle PK fractions during agarose bead gel
filtration. ‘

Figure 5. 280 nm Absorbance and smooth muscle PK fractions during elution from
carboxy-methyl agarose ion exchange column.

Figure 6. 280 nm Absorbance and smooth muscle PK fractions during elution from Blue
Dye - agarose column at pH 6.3.

Figure 7. 280 nm Absorbance and smooth muscle PK fractions during elution from Blue
Dye - agarose column at pH 7.5.

Figure 8. Capillary electropherogram of rabbit bladder metabolites in borate buffer.

Figure 9. Capillary electropherograms of rabbit bladder metabolites with additions of
NAD+, UDP-glucose, PCr, GTP, and NADH.

Figure 10. Reproducibility of the same rabbit bladder extract capillary
electropherograms.

Figure 11. Non-effect of CE peak height on peak height variability.

Figure 12. Comparisons of rabbit bladders metabolites from two different rabbit using
capillarly electrophoresis.

Figure 13. Variation in capillary electropherograms of rabbit bladder metabolites with
injection duration (and volume).

Figure 14. Change in selected rabbit bladder metabolite peak heights as a function of
injection duration.

viii

Figure 15. Influence of prior capillary electrophoresis injection duration on a 2-sec test
injection of rabbit bladder metabolites.

Figure 16. Temperature dependence of skeletal muscle pyruvate kinase absorbance.

Figure 17. Changes in capillary electropherograms of norepinephrine during ascorbic
acid additions in borate buffer at pH 7.4 and 9.4.

Figure 18. Electric field dependence of norepinephrine—ascorbic acid binding.

Figure 19. Linearization of NE-AA binding and determination of the binding constant
and 95% confidence interval in CB experiments.

Figure 20. Extrapolation of The NE-AA binding constant to zero electric field from the
electric field dependence of NE- AA binding.

Figure 21. Linearizations of NE-AA binding at different NE concentrations.

Figure 22. Concentration dependent change (fugacity) of NE-AA binding in an electric
field.

Figure 23. Effect of ascorbic acid - related compounds on CE binding to norepinephrine.
Figure 24. Effect of norepinephrine - related compounds on CE binding to ascorbic acid.

Figure 25. Extrapolation of NE—morphine binding to zero electric field from electric filed
dependence of NE-morphine binding.

Figure 26. Comparison of NE-AA and NE-morphine extrapolations of CE determined
binding constants.

Figure 27. Temporal development of absorbance change of combining MM creatine
kinase and skeletal muscle pyruvate kinase.

Figure 28. Spectrums of skeletal PK alone, MMCK alone, and a 0.5-0.5 mix of SKPK
and MMCK.

Figure 29. SDS gels of smooth muscle and skeletal muscle pyruvate kinase.
Figure 30. Capillary electropherogram of beef bladder extract.

Figure 31. CE-measured beef bladder metabolites from five bladders used for smooth
muscle PK isolation.

Figure 32. Lineweaver-Burk plots of ADP dependence of skeletal muscle PK.

ix

Figure 33. Lineweaver—Burk plots of ADP dependence of smooth muscle PK.

Figure 34. Comparison of the ADP-dependent kinetics of SKPK and SMPK.

Figure 35. Lineweaver—Burk plots of PEP dependence of skeletal muscle PK.

Figure 36. Lineweaver-Burk plots of PEP dependence of smooth muscle PK.

Figure 37. Comparison of the PEP-dependent kinetics of SKPK and SMPK.

Figure 38. Temperature dependence of SKPK and SMPK activity.

Figure 39. Activation energy determination of SKPK and SMPK.

Figure 40. 195 nm measurement of the expected absorbance of solutions containing both
PK and CK from the absorbance of individual solutions containing SKPK, SMPK or

MMCK.

Figure 41. Concentration dependence of the peak areas of capillary electropherograms of
solutions of MM creatine kinase.

Figure 42. Capillary electropherograms of MMCK, SKPK, and their combination.

Figure 43. Ratio of the CE absorbance of MMCK + SKPK to the absorbance of MMCK
as the creatine kinase concentration increases at 204 V/cm.

Figure 44. The CE absorbance difference between MMCK and MMCK + SKPK as the
MMCK concentration increases at 204 V/cm.

Figure 45. Comparison of the 340 nm absorbance differences produced by combinations
of MMCK + SKPK and BBCK + SMPK.

Figure 46. Effect of BBCK and ethanol on SMPK activity.
Figure 47. Effect of MMCK and ethanol on SKPK activity.
Figure 48. Effect of both MMCK and BBCK on the activity of SKPK and SMPK.

Figure 49. Effect of both MMCK and BBCK on the activity of SKPK and SMPK per unit
of PK.

Figure 50. Estimation of the electric field generated by a cell membrane.

Figure 51. Estimation of the change in NE-AA dissociation constant at different
distances from a cell membrane.

Figure 52. Experimental range of Ke measurements of NE-AA dissociation relative to
the electric field generated by a cell membrane.

xi

ADP
ATP
BBCK

CAMP
CDTA
CE

CK
CoA
DMSO
EOF
F-actin
FRAP
g

kD
Km
LDH
Mi-CK
MMCK
NAD+
N ADH
NE
NMR
PCr
PEP
PK
SDS—PAGE
SKPK
SMPK
TCA
TEMED
U

UTP
GTP
Vmax
w/v

LIST OF ABBREVIATIONS

ascorbic acid

ascorbic acid phosphate

adenosine diphosphate

adenosine triphosphate

BB isoform of CK

degree Celcius

cyclic adenosine triphosphate
trans-1,2-diaminocyclohexane-N,N,N ’ ,N’ -tetraacetic acid
capillary electrophoresis

creatine kinase

coenzyme A

dimethyl sulfoxide

electroosmotic flow

filamentous actin

fluorescence recovery after photobleaching

force of gravity

kilodalton

Michaelis-Menten constant

lactate dehydrogenase

mitochondrial CK

MM isoform of CK

nicotinamide adenine dinucleotide (oxidised form)
nicotinamide adenine dinucleotide (reduced form)
norepinephrine

nuclear magnetic resonance

phosphocreatine

phosphoenol pyruvate

pyruvate kinase

sodium dodecyl sulfate polyacrylamide gel electrophoresis
skeletal muscle pyruvate kinase

smooth muscle pyruvate kinase

tricarboxilic acid

N,N ,N’ ,N’ -tetramethylenediamine

enzyme activity unit

uridine triphosphate

guanosine triphosphate

maximum velocity constant of the Michaelis-Menten equation
weight per volume

xii

INTRODUCTION

The transfer of ATP between SKPK and MMCK was discovered in 1989
using NMR saturation techniques (Dillon and Clark, 1990). Since then many types
of functional and physical interaction have been discovered between these two
enzymes. These interactions were detected by changes in migration rates on paper
chromatography, in absorbance, in solubility, and in activity.

These interactions are important because they connect two areas of
metabolism which are closely related: energy production from glycolysis, the PK
reaction, and energy buffering, the CK reaction. The PK-CK interaction is the most
direct connection between the two. These interactions are also important on a more
general level because they are further evidence supporting the current trend in the
modern view of metabolism. This modern view attributes greater importance to
protein interactions than the classical view of metabolism.

These interactions were studied further in relation to the enzymes of smooth
muscle. This was done because smooth muscle has lower energy fluxes than skeletal
muscle, and its energy fluxes do no change as rapidly. Many of the features of the
modern view of metabolism are best illustrated as ways to tightly control energy
fluxes in tissues which have the ability to make rapid and large changes in these
fluxes.

The PK-CK interaction is an interaction which can occur while the enzymes
are free in solution. Capillary electrophoresis (CE) methods were developed in order
to study such interactions. Both PK and CK are difficult to run through capillary

tubes but in the course of these investigations a new method for measuring binding

1

of small molecules was developed. This method had advantages of low sample
requirements and quantitative results. More importantly this method yielded new
information about the binding of molecules in that the effect of an electric field on
the association of free molecules could be determined. This spurred the development
of new theories relating to the behavior of moleclues which bind to each other when
they are in proximity to membranes.

These new findings are not limited to molecules smaller than proteins as it is
possible to show the effect using antibodies. So far the quantitative method is
limited by the adherence of the molecules to the capillary. This is a problem with
kinases of intermediary metabolism and glycolytic enzymes in general. Despite
these problems it was possible to show a qualitative effect with SKPK and MMCK.
In addition, absorbance and activity techniques have been used to compare the

smooth muscle and skeletal muscle pyruvate and creatine kinases.

BACKGROUND

The classical view of metabolism.

Principles of Michaelis-Menton kinetics.

The classical view of metabolism assumes that most enzymes are free in
solution. Michaelian kinetics are the basis for the control of their activity. Although
the concentrations of substrates and products are manipulated in vitro in order to
determine the kinetic parameters of enzymes, namely the Kd and Vmax; the range of
concentrations found in the cell is restricted. This is because the total enzyme
activity in the cell is adjusted to meet the demands of the pathways to which they
belong. If the enzyme is not regulated, the Km of the enzyme is expected to be close
to the usual substrate concentration so that the enzyme is responsive to changes in
substrate concentration. An enzyme with a hyperbolic concentration curve would
also be responsive to changes in substrate concentration when this concentration was
much lower than the Km, but there is no benefit to having such a Km. Enzymes
which are Michaelian while free in dilute solutions but coupled because of their local
cellular environment would tend to have Km's and Vmax‘s which are not adjusted to

the cellular conditions.

Regulatory enzyme kinetics and similarities with the modern view concepts.

Several modifications of this basic pattern are able to increase the efficiency
of enzymes: cooperativity, allosteric regulation, and covalent modification.
Cooperativity is usually seen in multisubunit enzymes because permanently bound
subunits have a direct link to each other. The interaction of freely soluble enzymes
with each other is similar in that there is information transferred from one enzyme to
another about the activity in that particular enzyme's pathway. In cooperative
enzymes both parts are operating in the same metabolic pathway. When enzymes
from different pathways interact there is the possibility of coordinated control
between pathways. When freely soluble enzymes interact, there is a third possibility.
If some metabolites are common between the two pathways, it is then possible to
create a link between the two which can act as an additional control site. Like
cooperativety this is a way to control metabolic pathways more precisely.

In a sense, cooperativaty is a form of allosteric regulation, while channeling
of metabolites has the added feature of direct transfer of metabolites. In positive
cooperativity the enzyme is turned into an on-off switch by the sigmoidal
dependence of its rate on substrate concentration. Enzymes which channel
metabolites in a hand-off manner while they are not permanently bound can have a
similar property. In cases where there is no control of channeling, as in tryptophan
synthase, there is one multifunctional enzyme with several activities and no
intermediary is released. This type of channeling is an efficiency feature but it
cannot be controlled. Conversely, separate enzymes which act in a hand-off manner
can act as on-off switches if the coupling is itself controlled.

For cooperative enzymes low substrate concentrations cause the rate of
catalysis to be low. Unlike enzymes where the rate has a hyperbolic relationship to

the substrate concentration, the cooperative enzymes do not increase in rate quickly

4

as the substrate concentrations increase from these low levels. But around the
substrate concentration which causes half maximal activity, the change in rate with
respect to the change in substrate concentration--the second derivative of the
substrate concentration with respect to time--reaches its highest point. In this case
the enzyme may help maintain the substrate close to the substrate concentration at
which it causes half maximal activity. This is especially useful for substrates which
are used by several pathways but it cannot directly regulate which pathway uses the
metabolite. Regulated channeling does have this added control feature.

Two of the roles that the modern view of metabolism assigns to protein
interactions is the control of such branch points independently of this mechanism.
Substrates that have two competing roles in that they are intermediates for several
pathways are often regulators of key enzymes (Newsholme, 1981, p. 70). In a
similar way, the interactions that these enzymes have with enzymes of other
pathways may coordinate parallel pathways that have a common metabolite. The
sigmoid behavior is a response to this problem: a 10% to 90% change in activity on
a hyperbolic activity curve would require an 80 fold increase in substrate
concentration while the same change in activity can be achieved with a 4 fold
increase in substrate concentration by a sigmoid activity curve (Newsholme, 1981, p.
70-71). This makes it possible for the substrates to remain in the concentration range
needed by various pathways while allowing them to perform their regulatory
function. Channeling through coupled enzymes would also be useful in performing
this dual role; channeling may have a different function in enzymes which operate
far from equilibrium.

Allosteric regulation is also involved in the regulation of enzymes by
molecules binding to sites other than the catalytic center. In this case activation by
molecules earlier in the pathway or inhibition by molecules later in the pathway

results in the control of the enzyme to match the availability of substrate or need of

product. A well studied example is the binding of fructose 2,6-bisphosphate to
phosphofructokinase. Fructose 2,6-bisphosphate is formed from fructose 6-
phosphate early in glycolysis. When it binds to phosphofructokinase it causes a
conformational change and a resulting shift in the activity curve of
phosphofructokinase with respect to fructose 6-phosphate. The curve becomes
hyperbolic thus increasing the activity. These effectors can also work across
pathways as coordinating signals. The best example, ATP, is used by almost every
pathway and is also an allosteric effector in several pathways. It is allosteric on PK
and phosphofructokinase in glycolysis, on isocitrate dehydrogenase in the TCA, on
dihydroorotase in pyrimidine biosynthesis, on cholesterol monooxygenase in steroid
biosynthesis, on myo-inositol oxygenase in inositol metabolism, and others. The
skeletal PK is not effected in this way so channeling in this tissue is probably not
related to whole cell ATP control.

The need for tight regulation and quick response times comes into evidence
in several well characterized systems. Some pathways maintain elaborate series of
enzymes whose function is to create a signal amplification, to balance the rate of
competing pathways, or to decrease the response time of the pathway. A good
example which acts in all three capacities is the signaling pathway involving
epinephrine, CAMP, glycogen synthase, and glycogen phosphorylase. More
evidence for the importance of tight regulation and quick response times comes from
the existence of substrate cycles, sometimes called futile cycles. In these systems
enzymes in opposing pathways catalyze reactions which reverse each other's effect
on the metabolites in question while using up a form of energy such as ATP. The
product of one reaction is the substrate for the next and the product of the second
reaction is the substrate for the initial reaction. While one pathway is on there is a
low level of activation in the opposing pathway continually using energy. An

example is the continual phosphorylation and hydrolysis of fructose 6-phosphate.

6

Reciprocal allosteric control tends to make one pathway predominant but isotope
labeling has shown that during gluconeogenesis fructose 6—phosphate is still
phosphorylated (Katz and Rognstad, 1976). Such enzyme pairs may help amplify
metabolic signals (Newsholme et al., 1984) or decrease the response time of a switch
from one metabolic state to another (Newsholme and Start, 1981, pp. 71-73). These
two examples are important to the concept of channeling. They show that, although
a model of the cell's metabolism can be designed without them, these phenomena are
important enough in increasing efficiency that they exist despite their use of energy.
In the same way, channeling is not necessary for a coherent model of cell
metabolism but its existence is nevertheless present in some systems and should be
included in models of cell metabolism.

The value of a mechanism like a futile cycle can be estimated by its energy
expenditure at times when it is not in use. Regulated channeling can have the same

effect without incurring the maintenance cost.

Multi-functional enzymes and enzyme complexes.

Multi-functional enzymes and stable enzymes complexes give further levels
of control but do not alter the classical view significantly. The pyruvate
dehydrogenase complex is a large complex made of 60 peptides and three different
species of prosthetic groups. In four steps it catalyzes the conversion of pyruvate
into acetyl-CoA and all intermediates are tightly bound to the enzyme complex. In
such a case it would be impossible to regulate the individual subunits without
effecting the rate of the entire complex. Clearly, multi-functional enzymes and
enzyme complexes can be treated as single catalytic units. This type of channeling
does not let intermediates accumulate in the bulk phase and, as such, it is part of an

extended multi-step catalytic mechanism. Another type of channeling occurs in

7

tryptophan synthase. This enzyme catalyzes two sequential steps in the biosynthesis
of tryptophan: cleavage of indole-3-glycerol phosphate--which yields the indole
interrnediate--and the condensation of indole with serine. During the steady state
reaction, it has been impossible to trap the indole and this led to the conclusion that
channeling was occurring (Demoss, 1962; Matchett, 1974). The crystal structure
was found to contain a 25-A hydrophobic tunnel connecting the two catalytic sites
(Hyde et al., 1988). In 1994 Schlichting et al. made a mutant tryptophan synthase
with a bulkier amino acid in the tunnel. They showed transient formation of indole
and obstruction of the channel by rapid chemical quench methods and x-ray
chrystallography, thus confirming the suspected mechanism (Schlichting et al.,
1994). In this case it appears that the intermediate has to travel from one site to the
next but is still not able to escape into the bulk phase; it is still impossible to regulate

either the channeling or the individual steps.

Compartmentation within organelles.

Compartmentation within organelles establishes physical compartments in
which classical metabolism operates. This modifies the classical view further since
within a cell several pathways will have intermediates that are not available to each
others' enzymes, also a putative function of channeling. Organelles also maintain
environmental conditions or structural organization which are conducive to certain
pathways. For example the mitochondrial matrix maintains a pH and calcium
concentration far from the cytoplasm's. Also, the inner mitochondrial membrane is
the structural basis for the coupling of oxidative processes in the electron transport
system and for the phosphorylation of ADP by providing a hydrogen ion barrier.

Finally, organelles help keep intermediates and their enzymes in the same area. The

organelles then are a way of sub-dividing classical metabolic pathways and
providing the different structural and environmental necessities of these pathways.

It is interesting to note that functional compartmentation of reductive power
has been achieved without organelles by the use of the NADH/NADPH pair.
Glycolysis and biosynthesis are able to run at full speed concomitantly even though
they are in the same physical compartment. Classical channeling may achieve the
same result on a smaller scale but the linking of entire pathways in this manner is
unlikely.

Channeling among free enzymes and localization of pathways within areas of
a single compartment is intrinsically different from compartmentation in organelles
in that it supplies only some of the functions of organelle compartmentation but is
able to do it in a dynamic and regulated way: Changes in energy demand don't
translocate the TCA cycle to the cytoplasm but phosphofructokinase binding to sub-
cellular structures does correlate with changes in energy demand (Luther & Lee,
1986). Dynamic channeling would make it easier to coordinate the regulation of
diverse pathways because regulatory enzymes could act as the coordinators while
equilibrium enzymes would be affected by their location. Phosphofructokinase may
be able to coordinate glycolysis in this way by altering its binding to subcellular

structures and changing its activity (Marmillot et al., 1992).

Experiments on isolated enzymes as basis for classical view; its inadequacy.

The classical view of metabolism comes from the approach of most studies.
Enzymes are isolated from cells. Their activity is tested against various substrates
and effectors. Metabolic pathways are reconstructed from the behavior of the
individual enzymes. Finally, models of metabolism in cells are designed. As these

models get more complicated it becomes harder to predict the effect of alterations on

9

aspects of metabolism. Interactions between pathways which are unsuspected are
generally not tested and remain undiscovered until a perturbation in vivo did not
yield the theoretically expected result. The metabolism of Escherichia coli is one of
the most well characterized, yet overexpression of a few genes was found to effect
the expression of many unlinked genes through unexpected alterations in metabolism
(Liao et al., 1994). This is the largest problem in the field of genetic engineering
where modeling is now being extensively used (Bailey, 1991; Stephanopoulos and
Vallino, 1991). These problems are not necessarily due to channeling, but they
illustrate how our understanding of how pathways are integrated depends
increasingly on empirical evidence as we view more pathways together. This is a
prime concern to the field of metabolic engineering because its main concern is the
result of altering individual pathways in a system which does contain all the
pathways.

Certain phenomena such as metabolite oscillations are difficult to predict.
They are not expected from the classical theory unless mathematical modeling is
done. Modeling of substrate behavior in one dimensional (Ottaway and Mowbray,
1976, pp. 138-143) or two dimensional space may yield such information
(Marmillot et al., 1992), but modeling reactions in three dimensional space is much
more difficult and even two dimensional results do not necessarily apply in three
dimensions. Even when no new parameters arise the same values may yield
different results in three dimensions. Adding a time parameter to non-steady state
models renders many of these problems intractable. Sometimes anomalous
experimental findings in vivo can help clarify the relationship between pathways.
This is the case with oscillatory behavior. It is usually not present in vivo but can be
seen under certain conditions or as oscillatory relaxations after perturbations and can
bring new relationships to light (Kohen et al., 1987). Sometimes temporal modeling

gives oscillatory results. This has occurred with the pyruvate dehydrogenase

10

complex (Selvanov et al., 1994) and phosphofructokinase (Termonia and Ross,
1981). Another such phenomenon is the existence of substrate gradients; these have
been seen in experiments (Blaedel et al., 1972; Thomas et al., 1972). When modeled
these results are generally sensitive to the parameters which are difficult to measure.
This sensitivity makes these phenomena of possible use in understanding the
regulatory roles of substrates.

PK and CK have simple roles according to the classical view. PK catalyzes
the last step of glycolysis and is allosterically regulated. CK catalyzes the exchange
of phosphate between ADP and creatine thus maintaining a readily available pool of
phosphates for rephosphorylating ADP. But even without dynamic channeling, these
roles take on new levels of complexity when we analyze them in terms of entire

metabolic systems.

Classical view of PK.

PK is an important enzyme. It controls the flow of three carbon units arising
from glycolysis to pyruvate. Pyruvate has several fates. It can be oxidatively
decarboxylated to acetyl CoA to supply fuel to the TCA cycle. During anaerobic
glycolysis it can be used to replenish the NAD+ which was used earlier in glycolysis
and the lactate can be released into the blood. In one step it can also be converted
into alanine and from there to phenylalanine. More generally, glycolysis is the main
flux channel during times of acute high energy use. It then produces ATP
anaerobically. During prolonged high energy use it supplies the pyruvate which will
be converted to acetyl CoA. Glycolysis is also connected to other pathways: It
provides intermediates for the biosynthesis of several amino acids; through it's three
connections with the pentose phosphate pathway it supplies the right balance of

reductive power for biosyntheses and ribose units for purine and pyrimidine

ll

nucleotide synthesis; via dihydroxyacetone phosphate it can yield glycerol 3-
phosphate which enters lipid metabolism; it can supply intermediates to maintain the
capacity of the TCA cycle via PEP to oxaloacetate; and it is connected to numerous
peripheral pathways such as inositol metabolism via glucose 1-phosphate and
erythrocyte regulation via 2,3-bisphosphoglycerate. With such a central role in
metabolism there is hardly a time when glycolysis is not being closely regulated to
perform various conflicting tasks. As an integral part of the ATP production by
glycolysis and as a key regulatory enzyme of glycolytic outflow, PK is a critical
enzyme.

PK is a tetrameric enzyme with molecular weight approximately 237 kD.
There are several named PK isoforms: L in liver, M1 in skeletal muscle, M2 or A or
K in visceral organs, fetal, and neoplastic tissues, and R in erythrocytes. Cardenas
and Dyson (1973) have reported that the mammalian type L gives three
electrophoretically distinguishible isozymes and that type L isozyme can form
hybrids with type M from skeletal muscle. L-PK is the most regulated. It is
allosterically activated by fructose 1,6-bisphosphate and inhibited by ATP and
alanine. It also undergoes CAMP-dependent phosphorylation in response to
gluconeogenic hormones. This phosphorylated PK is less active. Ml-PK is the least
regulated of the PK isozymes and is not regulated by phosphorylation. It is not
allosterically affected by fructose 1,6-bisphosphate and is inhibited by phenylalanine
more than alanine. Kurganov et al. (1985) list fructose 1,6-bisphosphate, citrate in
low concentrations, and 6-phosphogluconate as activators; and acetyl CoA as an
inhibitor. M2-PK is like L-PK kinetically but is not inhibited by alanine. It has also
been shown to undergo phosphorylation in pancreatic islets but this does not appear
to have an effect on its activity (MacDonald and Chang, 1985)--a phenomenon
known as silent phosphorylation and that has been attributed to various functions

including regulation of degradation rates. Baranowska and Baranowski (1977) have

12

reported that 25% of PK activity from human skeletal muscle can be isolated in an
RNA-PK complex and that the RNA may have a regulatory function (1982) although
changes in activity due to complexing also occurred with heterologous RNA.
Oberfelder et al. (1984) showed that the phenylalanine binding causes a change in
conformation in Ml-PK which is 90% complete by the time 25% of the binding sites
are taken up. This was taken as evidence that Ml-PK exists in either of two
conformational states which can be described by the Monod-Wyman—Changeux
model first used to describe the oxyhemoglobin dissociation curve. Conformational
change and silent phosphorilation may have implications for the diazymatic
interaction or for the activity change which accompanies it.

The mechanism of the PK reaction is thought to be associative (Hasset et al.,
1982). This is done in a way analogous to an 8N2 reaction with the nucleophile
attacking the anti-bonding orbital of a carbon and forming a single transition state.
In the case of phosphate transfer the PEP phosphate would form its transition state
by coordinating bonds of three of its oxygens with PK in a plane, with the PEP
bonded oxygen also stabilized by PK on one side of the plane, and forming a fifth
bond with the ADP molecule on the other side of the plane (Hassett et al., 1982).

Classical view of CK and refinements from experimental results.

CK is also an important enzyme in that it helps maintain the energy charge
during short periods of high ATP use. The reaction it catalyzes is near equilibrium
under cellular conditions and it is therefore useful as a buffer. The change in ADP
concentration due to cellular activity is fractionally much greater than the change in
ATP concentration so that rapid rephosphorylation of ADP is able to maintain the
ATP concentration near its original value. CK is generally found in tissues that have

short and sudden energy demands which cannot be met by the adenylate kinase

13

reaction or the rephosphorylation of ATP during glycolysis. It is present in all
muscle types, in brain, in photoreceptors, and in spermatozoa, but is absent in liver
where the energy flux may be high but more continuous (W allimann et al., 1989). It
is a dimer and there are two types of subunits: M, muscle, and B, brain. MM-CK is
found in skeletal muscle and heart, BB-CK in the brain and heart, and both in
smooth muscle. MB-CK is the predominant form in heart (Van der Veen and
Willebrands, 1966) and small amounts have been found in smooth muscle (Clark,
1990). The M and B subunits are conserved; enzymatically active hybrid MB-CK
has been obtained by combining subunits from different mammalian species
(Perryman et al., 1983). There is a mitochondrial CK (Mi-CK) which has no

structural relationship to M-CK or B-CK. It is an integral membrane protein.

The modern view of metabolism, non-specific interactions.

Concentration and diffusion of proteins in the cellular environment.

The protein concentration inside cells is high. It has not been possible to
attain the cellular concentrations of some proteins in vitro. Substrates generally
diffuse 100 times faster than proteins in dilute solutions (Ottaway and Mowbray,
1976, p.137) and proteins diffuse about 10 to 100 times slower in cells than in
dilution solutions (Dick, 1963). Ottaway and Mowbray (1976) note that regardless
of whether or not the proteins are truly in solution, they act as stationary objects
compared to the substrates. Some proteins appear to have different mobilities in
different parts of the cell. This may have been the case with aldolase in Swiss 3T3

cells (Pagliaro, 1993).

14

With the application of fluorescence redistribution after photobleaching
(FRAP) diffusion rates of proteins can be measured. This technique involves
microinjection of trace amounts of a fluorescent protein into the cytoplasm of a cell.
A small area of the cell is monitored and the fluorescent probe in that region is
permanently bleached with a laser. The reappearance of fluorescent molecules is
measured and a diffusional constant can be calculated. Using this method, Brass et
al. (1986) showed that the diffusion coefficient of the maltose binding protein in E.
coli was one thousandth of that measured in water. Using this method Pagliaro
(1993) showed that the diffusion coefficient of aldolase-~which Was shown to bind
actin in vitro--was higher in the perinuclear area where actin content is lower. This
may be due to dynamic interactions on a time scale which cannot be resolved by
FRAP. Cellular diffusion of macromolecules may not only be affected by the non-
Newtonian properties of the medium; dynamic binding to sub-cellular structures may
cause the cytoplasmic matrix to act as a warped diffusional field.

It is thought that the cellular protein concentration is so high that it has a
significant effect by water exclusion on the substrate concentrations and on the ionic
environment (Ottaway and Mowbray, 1976). Type 11 water is water whose rotation
has been significantly impaired as measured in NMR experiments (Cooke and
Kuntz, 1974). Type H water--which surrounds proteins and is structurally more
organized by the generally anionic character of most proteins-~makes up 20% of
cellular water (Ottaway and Mowbray, 1976, p. 140). There is evidence that some
cells are able to control this and that this may be one mechanism for cell motility. A
high concentration of protein and their adjacent water molecules causes a high
degree of organization of water. Not only does this alter the dielectric constant of
water, but it makes it difficult to define on the scale of proteins since the electric
field around the protein is not homogeneous. This makes it difficult to

thermodynamically characterize protein interactions. For example, if the surfaces of

15

interacting proteins are known, it is still difficult to estimate the equilibrium constant
for the interaction because the solvation of the protein is altered and conformational
changes result in a transfer of energy to or from the internal structure of the protein.
Conformational changes can also lead to changes in the trapped water content of
proteins (Colombo et al., 1992). The additional effect of changes in internal water
molecules also makes thermodynamic calculations difficult if conformational

changes occur.

Water exclusion and true substrate concentration, diffusion of substrates.

If water exclusion is important this could have an effect on substrate
concentration. Red blood cell cytoplasm is about 35% w/w hemoglobin which leads
to a non-ideal solution where activities of substrates are far from their concentrations
(Creighton, 1984, p. 339). Thermodynamic equilibria in these conditions are
different from those expected by measuring equilibrium constants in dilute solutions
(Creighton, 1984, p. 339). Although the effect on proteins is greater than on
substrates, slower diffusion rates of these substrates and higher effective
concentrations may lead to an increased localization of substrates. Substrate
concentrations in vivo are usually lower than their Km's (Brooks and Storey, 1991, p.
372). This may signify that activities are higher than expected in the cell or it may
signify that the apparent activity--in terms of pathway fluxes--in the cell is lower
because of specific enzyme-enzyme interactions which reduce diffusional distances
or decrease the Km's in an allosteric manner. Also, calculations that show that
waves of substrate concentration need cells whose size is at least the average size of
mammalian cells (Marmillot et al., 1992) may underestimate the effect of decreasing

the diffusion rate.

16

Atkinson (1977) has noted that solvation capacity is a limiting factor and its
optimization may be a driving evolutionary principle. Since most substrates have
only one fate, channeling at intermediary steps would beneficial (Srere, 1987, p. 93).
Although the concept of optimization of solvation capacity as an evolutionary
driving force of leaky channeling has been criticized (Comish-Bowden, 1991), tight
channeling may still be used in the intermediate steps of a pathway. Even leaky
channeling will be a target of optimization if it has the effect of matching substrate

availability with the enzymes' various Km's.

Dynamic complexes in membranes.

It is interesting that the diffusion coefficients of proteins in membranes were
found to be larger than expected. Glycophorin, for example, has about 50 times
more mass than phospholipids tested but has a diffusion coefficient only half of the
phospholipids' (Sackmann, 1983, p. 435, col. 2). This has been attributed to the
weak dependence of lateral diffusion on the radius of the particle. The concept of
channeling between freely diffusing enzymes in the mitochondrial membrane is well
established. The search for super-complexes of glycolytic enzymes has not yielded
good evidence and the occurrence of super-complexes in the inner mitochondrial
membrane is not resolved. This membrane has conditions which would tend to favor
complex formation: 50% of the surface is protein and 35% of the protein in the
membrane is involved in a single pathway--the electron transport system.
Hackenbrok (1981) has determined that the rate of diffusion of complexes in isolated
membranes is fast enough to account for the activity of the enzymes. These
experiments have been criticized because of the low viscosity of his system

compared to that found in vivo (Srere, 1987).

17

The modern view of metabolism, specific interactions.

Channeling of substrates by enzymes in solution, the diazymatic mechanism.

Apart from the channeling of substrates within enzyme complexes, there may
be many enzymes which channel in a hand-off manner, passing substrates by
temporarily binding in solution. This effect would be important if regulation of local
concentrations of metabolites occurs in a pathway. Brooks and Storey (1991, p. 373)
have noted that direct transfer between glycolytic enzymes is unlikely because of the
constraints imposed in lining up large multi-subunit enzymes. But as seen with
cytochrome c in two dimensions this may not always be the case. Large molecules
have a large number of electrostatic charges which may be involved in approaching
each other with the correct alignment and this may offset the problem of their mass.
Also the translational diffusion of proteins is probably slow enough compared to
their rotational diffusion that when two proteins collide they encounter a larger area
than the original collision area. Multi subunit proteins with identical subunits and
binding sites facing out also have a larger percentage surface area available for
binding at that site than the single subunits. Brooks and Storey also noted that these
interactions would be limited to pairs of enzymes which use substrates that undergo
cycles such as ATP and ADP or NADH and NAD+. But all enzymes using these
substrates still use other substrates which do not undergo cycles so their argument
would limit channeling to a smaller subset where the cycling intermediates are
tightly bound unless there is a channeling interaction. They also refute Srivastava
and Bemhard's (1988) hypothesis that it is the large off-rate constants of the coupling
reactions that make it difficult to see these complexes in dilute solutions by pointing
out that the overall flux through the glycolytic pathway cannot be changed by

increasing the number of interactions if the overall bound proportion of enzymes

18

does not change. These two arguments ignore allosteric behavior during protein-
protein coupling--a central feature of coupling theory. Although there are not many
pure cycling enzymes CK is an exception. CK catalyzes a reaction which acts as a
buffering mechanism so it could easily be used with a single other enzyme species.
Creatine would never have to leave CK if CK acted as a shuttle.

Channeling enzymes involved in unidirectional substrate flow could not
come together productively more often than specified by the product release rate of
the first enzyme. One way for channeling with free enzymes to be possible during
unidirectional substrate flow is for the first enzyme to have a relatively tightly bound
product. This would lengthen the time during which a successful encounter could
occur with the second enzyme. If the change in free energy due to the binding of the
enzymes was negative when the first enzyme is bound to its product, some of this
energy could be transferred to the first enzyme. This could lead to a conformational
change which would cause the release of the substrate. The substrate could travel
through the inter-protein space to the second enzyme and the free energy drop of its
binding there could cause this enzyme to alter its conformation to one less likely to
bind the first enzyme. Catalysis would then cause the drop in free energy necessary
for release of the final product from the second enzyme. Also the first enzyme
would bind another substrate and catalyze its reaction and the free energy drop from
that reaction would be conserved in the maintenance of the enzyme product complex.
Here the enzymes are ready to go again. The free energy of catalysis of the first
enzyme supplied the energy needed for the conformational change which made that
enzyme more likely to bind to the second enzyme. This process may be considered
an "energy transfer" scheme whereby the activation energy for each step is derived
directly from the previous step instead of through a bulk thermal mechanism

(Gaertner, 1978).

19

The "ambiquitous" model and localization of pathways.

Some enzymes bind to subcellular structures and have altered activity when
bound. Such enzymes have been called "ambiquitous" (Wilson, 1980). The best
example is phosphofructokinase which was shown to bind to F-actin and have an
increase in activity of 34% (Luther & Lee, 1986). Other enzymes which have been
shown to bind to F-actin are lactate dehydrogenase, phosphoglycerate kinase, and
triosephosphateisomerase (Brooks and Storey, 1991, p. 362). Using FRAP, Pagliaro
(1993) showed that aldolase had a 20% bound fraction in the perinuclear area of
Swiss 3T3 cells while enolase had no bound fraction. This binding disappeared
when 2-deoxyglucose was put in the supporting medium. Since this 2-deoxyglucose
depletes glycolytic intermediates reaching aldolase, it implicates a dynamic
interaction between binding to subcellular structures and catalysis.

The binding of enzymes to subcellular structures also provides a way to
impart them with topographic information. This property, without which differential
localization of pathways is impossible, is absent in totally free molecules.
Localization of substrates may have any of a number of roles. Selective control of
substrate availability to various pathways is the main role. The following have also
been proposed: dynamic compartmentation of metabolites, signal transmission,
efficiency, excitability, metabolic-shift efficacy, and spurious kinetic phenomena

which are internally damped (Marmillot et al., 1992, p. 12106, top col. 2).

Metabolic pathway requirements: Matching Km's and isozymes.

Since it may not be valid to average substrate concentrations across the cell,
the Km's may yield additional information about particular pathways. One reason

for channeling and localization of pathways may be to regulate substrate

20

concentrations to levels which are optimized for the particular pathway. This would
explain the substrate concentrations being generally lower than the Km's (Brooks
and Storey, 1991) because it would be necessary for enzymes to be less active when
not interacting with a particular pathway. Also, different isoforms of one enzyme
might be used to keep substrates tracked to one pathway as opposed to another
competing pathway or to control pathways differentially depending on the metabolic
status of the cell (Ureta, 1991). One set of experiments have yielded positive results.
Cellular micro-injection of anti-hexokinase-C antibody inhibited 14C-glucose
incorporation into glycogen without stopping C02 output but 14C-glucose-6-
phosphate injection reversed the inhibition (Ureta, 1991). It may then be important
that smooth muscle contains both MM-CK and BB-CK. And does smooth muscle
contain several PK isozymes? Chicken liver has been shown to contain both L-PK
and M2-PK (Eigenbrodt and Schoner, 1977) so it appears that there are tissues with
several PK isozymes. Also, the silent phosphorylation of M2—PK may be a

mechanism by which it is transferred from one pathway to another.

Glycolysis and the modem view of metabolism.

The classical view, glycolysis, and channeling.

Identifying non-equilibrium reactions may be done by measuring the
maximal enzyme activity and the mass-action ratios (Newsholme and Start, 1981, p.
96). In analyzing glycolytic enzymes from various tissues it has been noted that the
relative activities of hexokinase, phosphofructokinase, aldolase, and enolase are low
and these enzymes are thus classified as non-equilibrium enzymes (Newsholme and

Start, 1981 ). However only hexokinase and phosphofructokinase are non-

21

equilibrium enzymes by their mass-action ratios (Newsholme and Start, 1981). The
high mass-action ratios do not suggest that channeling is occurring in these cases.
Even if glycolysis operated via channeling in most cases, there are definite cases of
uncoupling of specific energy production and energy use pathways, so channeling is
not excluded. This was noted in Paul's ( 1983) work when removal of glucose from
the medium didn't cause the contracture usually associated with inhibition of the
sodium-potassium ATPase but did result in a decrease in lactate production. It is
still difficult to tell when and where coupling is important. All pathways may be
coupled and regulated according to demand, or some may be coupled and
functioning at a high steady state while the overall uncoupled capacity is only being
used at a low level. In the later case, either increased coupling or increased
uncoupled flux could occur in times of stress. Enzymatic biochemical analyses in
vitro tell us that almost all systems are able to function in the uncoupled state. The
question of efficiency has been addressed by some in order to determine if the
activities measured in these conditions can account for the rates measured in vivo and
it has been concluded that many in vitro rates do not meet this requirement (Srere,
1987). Low relative activities of enzymes that are catalyzing equilibrium reactions
suggest that channeling may be occurring. In this case channeling would function as
a form of steady state flux enhancement and would maintain reduced transition times

of metabolites (Melendez-Hevia and Montero, 1991 ).

Glycolysis and the ambiquitous model.

Much work has been done on the binding properties of glycolytic enzymes
since the first associations with F-actin were studied in 1971. Brooks and Storey
(1991) have taken the association constants of the enzymes with each other and with

F-actin, and the activity changes reported for these enzymes from twenty-three

22

separate studies. From these values they designed a model of glycolysis. In this
model, they found that a substantial proportion of the enzymes were bound to either
F-actin or other enzymes. They also found that the long sought single glycolytic
complex did not exist. They found that the binary enzyme complexes increased the
flux through some enzymes but decreased it through others. And finally they found
that the binding of phosphofructokinase to F-actin could play a significant role in
regulating glycolytic flux.

It can be argued that slowly diffusing proteins cannot increase the flux of
quickly diffusing metabolites through pathways. Since the rate of catalysis is
generally slower than the diffusion rate it has even been argued that channeling
within complexes is unnecessary. Several analyses have shown that channeling can
in fact increase flux (Welch and Easterby, 1994; Smolen and Keizer, 1990). In the
case of unequal distribution of enzymes, it is the spatial distribution of the substrate
that is deemed important (Ottaway and Mowbray, 1976) and this was the explanation
used for the increase in catalysis seen by Mosbach (Srere et al., 1973) in his

experiments with glycolytic enzymes bound to a solid support.

Glycolytic channeling and ambiquity in vivo.

Physical coupling was also correlated in vivo with functional changes in
metabolism. In skeletal muscle, enzyme binding correlated with changes in
glycolytic flux brought on by exercise (Clarke et al., 1984; Brooks and Storey,
1988), and in cardiac muscle, binding occurred in response to ischemia (Choate et
al., 1985; Stephan et al., 1986).

The most striking example of channeling in glycolysis is R. Paul's work on
smooth muscle (1983). He recorded lactate production and oxygen consumption as

measures of glycolysis and oxidative phosphorylation respectively. He was

23

investigating the unusual aerobic glycolysis known to occur in smooth muscle (Paul
et al., 1979). When he stopped Na+-K"' transport with ouabain, lactate production
was inhibited. When he increased potassium influx, lactate production was
increased. In both cases oxygen consumption was found to parallel changes in
isometric force and ouabain had no affect on oxygen consumption even though
isometric force was maintained. It appears that there is functional coupling between
the Na+-K+ ATPase-pump with glycolysis on one hand; and coupling of oxidative
phosphorylation with force development on the other. There have been other reports
of coupling of glycolysis with membrane transport of sodium and potassium in
erythrocytes (Solomon, 1978) and in ascites tumor cells (Racker, 1976), and of
calcium in the sarcoplasmic reticulum (Entman et al., 1976). In smooth muscle,
glycolysis was associated with calcium uptake and both PK and CK were found in

membrane fractions (Hardin, 1992).

24

METHODS

Urinary bladder smooth muscle PK isolation.

Introduction.

Two procedures were followed for isolating PK from smooth muscle. The PK
used to determine the kinetic parameters was isolated using the first procedure.
Subsequently a more rapid second procedure was developed. The PK used in the CE
experiments and in the PK-CK interactions experiments was isolated using this second

procedure.

Bulk methods for the first procedure.

Early bulk techniques where adapted from Eigenbrodt and Schoner, 1977, and
altered in the course of developing a procedure specific for smooth muscle PK. Beef
urinary bladders were obtained from Bain's Packing and Refrigeration at the time of
slaughter and individually frozen in liquid nitrogen. When not mentioned otherwise, the
isolation was done in a cold room at 4 to 7 C and the centrifugation was at 25,000 g for
30 minutes.

The bladders were collected at the time of slaughter. Each bladder was removed
with scissors within five minutes of the kill. The fatty tissue and fascia were removed.
The bladder was cut open and rinsed in starting buffer cooled with ice. The bladder was

then packed in ice. The bladders were in ice for about five hours before the isolation was

25

started. The urothelium was first removed from the bladders with scissors and the
bladders were cut into small strips. The bladder smooth muscle and urothelium were
weighed separately and the smooth muscle was put in 1 ml/g-bladder of 100 mM Tris pH
7.25 and stirred for one hour at room temperature. The contents were then transferred to
a blender, cooled to 4 C, and homogenized for a total of fifteen minutes. This was done
in one minute intervals with the temperature monitored and cooling intervals with ice
bags were used as necessary. The contents were then divided into batches and each batch
done in sequence. Each batch was divided and centrifuged at 20,000 g for twenty
minutes. The precipitate was tested for activity but no re-homogenization was done
although as much as 15% of the activity remained in the precipitate. 100 ml of ethanol
per 1000 ml of supernatant was added under stirring and the mixture incubated for two
hours at 40 C. The denatured protein was removed by centrifugation. One ml of acetone
pre-cooled to -20 C per 2 ml of supernatant was added dropwise with stirring. This was
followed by centrifugation and lyophilization. The acetone step sometimes resulted in a
large fraction of the PK being permanently denatured. Bringing up the precipitated
protein in 100 mM imidazole, pH 7.6 yielded little activity even if the mixture was not

centrifuged. If the loss of activity was greater than 85% the isolation was restarted.

Bulk methods for the second procedure.

In this case the bladders were put on ice and kept that way until homogenization.
The homogenization was carried out as in the first procedure except that 100 mM
imidazole pH 7.6 was used. It had been determined earlier that 15% PEG precipitated PK
without precipitating most other proteins. Unfortunately most of the activity did not
appear in the brought up precipitate. Early experiments with PEG also caused much of
the PK brought back up to re-precipitate upon centrifugation. This second centrifugation

was necessary to clear the solution before the chromatography steps. By bringing up the

26

first precipitate to between 2 and 6 mg/ml of total protein and by mixing for several hours
at room temperature all the activity was retained.

Samples were taken from the supernatant of the first centrifugation. To each
sample was added an equal volume of PEG at a particular % w/v and the samples were
centrifuged. The precipitate was brought up with 100 mM imidazole, diammonium
citrate, or Tris. The original 100 mM imidazole, pH 7.6, was found to give an increase in
specific activity at a final nominal PEG concentration between 12% and 20% with the
best increase in specific activity at 12% (figure 1). Tris used at several pH's above 8 and
diammonium citrate at pH's 4.0, 4.5, and 5.0 caused much of the activity to remain in the
precipitate.

A 30% PEG solution was added in equal volume to the supernatant from the first
centrifugation. The PEG had an average molecular weight of 3,350. PEG with a
molecular weight of 400 was found to be ineffectual up to a concentration of 60%. The
post-PEG mixture was stirred for one hour at room temperature and then centrifuged.
The precipitate was brought up using 100 mM imidazole, pH 7.6, and centrifuged a
second time. The activity and total protein of the brought up precipitate and the
supernatant to the second centrifugation were measured. Pre-cent and post-cent in figure
3 refer to the second post-PEG centrifugation. The supernatant of the second
centrifugation was kept and was ready for the chromatography.

Only gel filtration and lyophilization was used to change the buffers for
experiments which required a specific buffer because of the high losses of activity
occurring during dialysis of small samples. Some of the lost activity could be recovered

from the dialysis membranes with a 500 mM NaCl wash.

27

.0
on

 

+ Precipitate

 

 

 

U/mg/ml sample)
0
U1

 

 

0-4 ' "k" Supernatant
V 0.3
.é‘
'5" 02 - 1
2 ' . 4,.
X _ I» " ‘ ,
Q_ 0.1
2 - 1“ ‘
:32 0,0 - tu-t-"t-"t-"k"
§ .-
Q'O.1 -
U) 1 . 4

 

O 5 1O .15 20 25 30
PEG Concentration (°/ow/v)

Figure 1. Effect of polyethylene glycol concentration on smooth muscle pyruvate kinase
activity.

28

Preparation for chromatography for the first procedure.

The precipitate was re-dissolved in 400 ml of 50 mM Tris, pH 7.5, and stirred at
room temperature for 3 hours. It was then incubated for 40 minutes in 50 C and
centrifuged. 10 m1 of 1 M MgClz and 10 ml of diethyleneglycol was added per 100 ml of
supernatant. This mixture was incubated for 2.5 hours at 45 C and centrifuged. The
precipitate was then taken up in 100 ml of 10 mM magnesium citrate, pH 8.5, and stirred
for 2 hours and centrifuged. The supernatant was then dialyzed for 12 hours against a
solution of 5 mM CDTA, 10 mM MgHPO4, 1 mM H3PO4, and 1 mM dithiothreotol
with the pH maintained at 4.8 with 1 N H3PO4. The solution was then centrifuged and
the precipitate stirred for 2 hours with 200 ml of 10 mM magnesium citrate, and 40 mM
sucrose, pH 8.5. The solution was centrifuged and the supernatant kept. The turbidity in
the solution was found to not interfere with the chromatography so no further bulk

methods were used.

Column chromatography for the first procedure.

First a hydroxyapatite column was used. Column equilibration was done with 1
mM NaH2P04, pH 6.8. The sample was prepared with the equilibration buffer and
glycerol was added to a final concentration of 1%. The proteins were eluted with 5 mM
MgC12, then a 1-500 mM NaH2P04 gradient. PK was eluted with the gradient.

The isolation sample was then loaded onto a carboxy-methyl cation exchange
column with loading buffer of 20 mM diammonium citrate, pH 4.8. PK was eluted using
a salt gradient of 0-1 M NaCl in 20 mM diammonium citrate, pH 4.8.

Further purification was achieved using a Bio-Rad affi-gel blue gel column. The
PK was loaded using 20 mM diammonium citrate, pH 4.8 and only bound partially to the

column but one protein peak was found to not bind at all. This second peak eluted early.

29

This step resulted in very dilute solutions. It was possible to prevent this dilution by
using a NaCl gradient (0-1M). After the sample was loaded, 5 ml were allowed to elute

and then the gradient was started.

CM-cation exchange experiment (second procedure).

This experiment gave the conditions necessary for the CM-cation exchange
chromatography to be consistent. The gel filtration step does not require a particular salt
concentration or pH so the conditions determined to cause tight binding of PK to the CM
groups were used during the gel filtration step. The gel filtration step tends to dilute the
sample and it is placed before the CM step. Since the CM group binds tightly to PK
under the appropriate conditions, the CM step can accommodate the large volumes of
sample created by the gel filtration step and concentrate the PK upon elution.

Ten 1.5 ml centrifuge tubes were each filled with 200 111 of CM-agarose beads.
These were washed repeatedly with 10 mM buffers at pH's evenly distributed between 4.0
and 8.5. When the pH of the collected wash no longer changed the tubes were filled with
buffer to a total volume of 1 ml. Then 50 ul of isolation sample was added to each tube.
The sample added was from a supernatant of the second centrifugation after the PEG
precipitation of PK. Enough sample had to be added to be able to measure the activity of
the sample after its dilution into the tube. Because the isolation sample had a higher
concentration buffer (100 mM imidazole) the pH in the tubes changed. The new pH was
estimated by mixing larger volumes of the appropriate buffers and measuring the new pH.
The tubes were shaken and the beads were allowed to settle. The activity was measured
and compared to the expected activity if there was no PK binding. Figure 2 shows the pH
dependence of PK binding to the CM groups. To determine if the binding caused loss of
activity, 20 ul of 1N NaOH was added to the pH 4.7 tube and the activity was re-

measured and compared to the expected activity. No loss of activity was expected from

30

0.022 f
0.020 j
0.018 j
0.016 f
0.014 f
0.012 f
0.010 j
0.008 j
0.006 T
0.004 j
0.002 j
0.000 '

umoles/3 ml/min +/- sd

 

4.5 5.0 5.5 6.0 6.5 7.0 7.5
“pH

Figure 2. pH - dependence of smooth muscle pyruvate kinase binding to carboxy-methyl
coated agarose beads.

31

adding 1N NaOH because it was added to the supernatant and was diluted before it
reached the PK bound on the beads. Figure 3 shows the maintenance of activity in
relation to the previous two centrifugations. The post-CM measurements were done after
the sample was added to the beads. The first post-CM is for a tube that did not bind PK
and shows that there is no loss of activity from the dilution in the new buffer, contact with
the beads, or contact with the plastic walls of the tube at this dilution. The second post-
CM, marked as post-pH, was the measurement from the pH 4.7 tube after the NaOH was

added. It shows that all the activity was recovered.

Column chromatography for the second procedure.

The chromatography for the second procedure, with the exception of the affinity
chromatography, was carried out at 25 C. The affinity chromatography was still carried
out at 4 C. The post-PEG sample was put through an agarose gel filtration column. In
figure 4 the PK can be seen to be at the head of the protein peak and the first three
fractions are kept for their high specific activity. The buffer used for elution was 10 mM
diammonium citrate, pH 5.3. The second peak in figure 4 was confirmed to be imidazole
by running several fractions on the CE. On the CE runs the imidazole peak height
matched the level of absorbance in the second peak in figure 4. The eluting PK was then
loaded onto a CM-cation exchange column using the same buffer. After two column
volumes of buffer have eluted a NaCl gradient is begun (0 to 1 M) and PK elutes at
approximately 300 mM N aCl (figure 5).

The PK was then loaded onto a Bio-Rad affi-gel blue gel column using 10 mM
diammonium citrate, pH 5.3 and bound partially to the column as in the first procedure
but one protein peak was found to not bind at all (figure 6). This second peak eluted
early. In this first large peak with no PK activity appears to have several peaks but this is

an artifact caused by protein that accumulated on the cuvette walls. Washing the cuvette

32

tended to drop the absorbance which would then rise again. As in the first procedure, this
step resulted in very dilute solutions. It was possible to prevent this dilution by using a
NaCl gradient (O-lM). After the sample was loaded, 5 ml were allowed to elute and
then the gradient was started. In earlier experiments, the PK peak was found to have a
low specific activity shoulder when a higher pH was used. This shoulder disappeared
when the gradient was added at the lower pH. A second run at a higher pH also showed a

low specific activity shoulder and was used as the final step in the isolation (figure 7).

Analysis methods.

The SDS-PAGE gels showing the major proteins at several steps and comparing
them to SKPK and MMCK are discussed in the results section. The gels were 20 cm
wide. The resolving gel was about 14 cm tall and the stacking gel was about 2 cm tall.
The procedures followed for polymerizing, loading, and running the gels were the Bio-
Rad procedures for the Protein ii electrophoresis system. The following conditions were
specific to the gels presented here. The polymerization was carried out exactly as the
instructions recommended except for the last three gels. Because of the age of the
chemicals, these were polymerized with 5% extra TEMED and ammonium persulfate.
The thin gels were 0.8 mm thick and run at 13 mA per gel during stacking and 19 mA
during resolving. The thick gels were 1.75 mm thick and run at 25 mA during stacking
and 40 mA during resolving. All were run under constant current conditions. They were
usually stained with coomassie brilliant blue. Some samples did not respond well to
lyophilization. Especially when the sample was small, much of the protein adhered to the
glass and was lost. After lyophilization, some buffers caused the proteins to clump and
were also lost. In these situations, when less protein was seen on the gels than expected,
it was possible to silver stain the gels after they had already been stained with coomassie

brilliant blue.

33

 

9 9: ?:9 9:99999999
«99«9«9999 «««9 99999«99 «9«9«9«9«9«9
vm«9m««99 9.9..9«99«9 99«9«9«««99«999.

9’99«««.<9 9999«9«« «««9«9«

 

 

 

\..\Q\§\\\\\

 

 

.///,/ /////

 

 

 

 

 

 

 

0.020 '

0.018 ’

086420
WMQ100000
00.000.00.00
000000000

3 .\+ 5:58 $8.85

Pre-Cent. ----'--------Post Cent.------------

------Post-CM------

Post pH

Figure 3. Non-effects of centrifugation, carboxy-methyl binding, and pH on smooth
34

muscle PK activity during SMPK isolation.

Absorbance (280 nm)
8
O
O

 

“'0— Absorbance
--A-- PK Activity

 

 

 

 

 

J

 

 

 

0 10 20 30 40 50 60 70 80 90100

fraction number

g

0.012
0.011
0.010
0.009
0.008
0.007
0.006
0.005
0.004
0.003
0.002
0.001
0.000

umoles/3ml/min

Figure 4. 280 nm Absorbance and smooth muscle PK fractions during agarose bead gel

filtration.

35

  
  
   

0.4 '

0.3 ’

0.2 '

 

 

 

 

Absorbance (280 nm)

0.1 '

‘
o.o “

 

 

 

'0' Absorbance ‘

- 1" PK Activity

 

 

 

 

25 27 29 31 33 35 37 39 41

ml

0.012
0.011
0.010
0.009
0.008
0.007
0.006
0.005
0.004
0.003
0.002
0.001
0.000

U/microliter

Figure 5. 280 nm Absorbance and smooth muscle PK fractions during elution from
carboxy-methyl agarose ion exchange column.

36

15'
13'

11'

 

  

9 i -— Abs
7 ~ """ PKACI

1 . I

g

 

 

 

 

Absorbance (280 nm)

 

    

I

I

l

1

1

I

1

1

1

1

1

1

1

1

1

1 \
1

........ a

 

 

 

o 10 20 30 4o 50 60 7o 80 90100
fraction number

‘ 2.5

2.0

1.5

1.0

0.5

‘ 0.0

Normalized PK Activity

Figure 6. 280 nm Absorbance and smooth muscle PK fractions during elution from Blue

Dye - agarose column at pH 6.3.

37

 

6.-
.. s-
E .
C
s 4'
a .
8 3
C
S
a 2-
(D .
2
1.
0.-
L

 

 

 

 

 

 

 

25 27 29 31 ""33 35 37 39

fraction number

‘ 1.4
1.2
‘ 1.0
‘ 0.8
' 0.6
‘ 0.4
‘ 0.2

‘ 0.0

Normalized PK Activity

Figure 7. 280 nm Absorbance and smooth muscle PK fractions during elution from Blue

Dye - agarose column at pH 7 .5.

38

During the bulk procedures the total protein was measured using the Bio-Rad
Bradford and De assays. When imidazole was used as the sample buffer during a Dc
assay it caused a white turbidity to appear in the tubes. The absorbance increase due to
the imidazole when no protein was present was determined. It was found that subtraction
of this expected absorbance from the total absorbance of a Dc assay of PK in an
imidazole buffer did not yield the absorbance due to the PK alone. For this reason the Dc
assay was not used in steps where PK was in imidazole. During the chromatography the
total protein was determined using the absorbance at 280 nm and the extinction
coefficient of PK (5.4 for 1% solution, 1cm path length). The absorbance at 280 was also
used to standardize the protein solutions from skeletal muscle.

Several PK assays were used. All assay were done using 50 mM imidazole pH
7.6, 240 uM NADH, 1.5 mM ADP, 1.5 mM PEP, 120 mM KCl, 62 mM MgSO4, 0.0135
U LDH. For the purpose of comparing the activity to literature values, the enzyme assay
was done at 37 C in 3 ml. For the purpose of comparing activity values within an
experiment, the assays were done at 25 C in 3 or 1 ml. The activity was measured by
following the decrease in absorbance at 340 nm over time as NADH was oxidized. A
unit is defined as one u-mole of PEP dephosphorylated per minute in the 3 m1 assay. For
the 1 ml assays the activity was converted to u-moles/min/3 ml so that the activities
would be comparable. The substrate solutions were run on the CE and their quality could
be monitored by observing the change in the ratio of the several peaks. For example, the
imidazole peak never changed and ran at the EOF. The NADH peak could be compared
to the imidazole peak and to the NAD+ peak. Also two small peaks appeared next to the
NADH peak and may be inhibitors which are mentioned in the Sigma Chemical Co.
information.

All the enzymes apart from the SMPK and all the substrates for enzyme assays
and CE standardization with one exception were bought from Sigma Chemical Co.

Phenylalanine was bought from Aldrich Chemical Co.

39

General CE methods used in all sections.

The conditions for all the runs are as follows unless stated otherwise. The
capillary had a diameter of 100 um, an overall length of 98 cm, and a length to the
window of 66 cm . The background buffer was 25 mM sodium borate, pH 9, for
experiments analyzing metabolites from bladders and assay substrates. The injection
time was 2 s for the metabolites and 5 s for the proteins. The absorbance range measured
was zero to 0.05 and the auto-zero was always done at or below the absorbance range
used. The rise time for the absorbance measurement was 3.2 s. The wavelength was 195
nm. The runs were carried out at a constant voltage of 20.0 kV. The current was
monitored and was approximately 90 uA. The current was proportionally lower in those

experiments where the voltage was reduced to change the electric field.

Bladder metabolites.

Rabbit bladder method.

Figure 8 shows a typical electropherogram of rabbit bladder metabolites. The
peaks were identified by spiking the samples with a known amount of one of the
metabolites. The peaks identified using this method were creatine, NAD+, lactate, UDP-
glucose, NADH, PCr, ADP, ATP, GTP, and UTP. The vehicle shows the effect of the
solution that the metabolites are extracted in and the background buffer in the capillary on
the absorbance. This includes a buffer peak which runs close to the electro-osmotic flow
position and a late peak also seen upon injection with no sample. Figure 9 shows

capillary electropherograms spiked with NAD+, UDP-glucose, PCr, GTP, and NADH.

40

The rabbit metabolite capillary electropherograms were reproducible enough to
give quantitative results without calculating the ratio of the metabolite peak to a standard
peak. When a standard peak is used it is placed in the sample buffer. Some buffers such
as imidazole can be used as a standard peak at most wavelengths. DMSO is a useful
standard peak in that it can be controlled independently of the buffer and runs at the EOF
position thus making it possible to accurately calculate the electrophoretic mobility of any
molecule in the run. Figure 10 shows the peak heights for many of the distinct peaks in
figure 9 and their standard deviations. Figure 11 shows that the standard deviation was a
constant fraction of mean peak height, about 0.10, and that it was not related to the peak
height. Figure 12 shows that the technique was accurate enough to measure differences
in bladder metabolites between two rabbits. Four of the peaks marked with an asterisk
were significantly different in the two rabbits, while the vast majority were not different.

When doing the CE runs the condition of the capillary may be effected by the
sample injection or the prior runs and so a procedure must be established which maintains
the runs identical. One effect is shown in figure 13. Note that as the injection duration
increases the migration times also increase. At the high injection durations the expected
peak height drops. The expected peak height in figure 13 was made the same for each run
by dropping the sensitivity as the amount injected was increased, yet the measured peak
height decreased noticeably in the last run. This effect is not statistically significant until
a 20 sec injection is used (figure 14).

The effect of the previous runs is complex. Sample injection tends to de-
condition the capillary wall and thus effect EOF. Figure 15 shows this effect. The
abscissa positions eight peaks according to their position during a run with a 2 sec.
control injection and following the conditioning of the capillary with 0.1 N NaOH. The
ordinate marks the ratio of the migration time of a 2 sec. injection done immediately after
a run with and injection time of either 2, 20, or 50 sec. to the 2 sec. control injections.

This clearly shows that the run times of compounds with a low overall velocity are much

41

Veh1clc

UTP

8..

ATP

 

 

ADP

     

NADH

   

UDPG

 

Lactate
NADl\ I

 

10 mm

 

 

 

VchIclc 9
C rcatmc

 

Figure 8. Capillary electropherogram of rabbit bladder metabolites in borate buffer.

42

(3)

A. Rabbit Bladder Extract
+ 0.042 mM NAD‘ (11 (1)
+ 0.006 mM UDPG (2)
+ 0.217 mM PCr (3)

      
  

 

 

WWW/1‘ 1

B. A + 0.032 mM GTP (4)

 

  

 

 

 

mm

Figure 9. Capillary electropherograms of rabbit bladder metabolites with additions of
NAD+, UDP-glucose, PCr, GTP, and NADH.

43

0.006 '

0.005 ’

0.004 *

 

 

 

 

0.003 '

abs +/- sd

0.002 '

 

 

0.001 - . 1

0.0001 1 11111 .. J1- .

8.3 13.6 19.0 24.4 29.7 35.0

 

 

 

 

min

Figure 10. Reproducibility of the same rabbit bladder extract capillary
electropherograms.

1

40.4

0.20 f
0.18 : '
0.16 : $ '

0.14 T
0.12j .
0.10 W
0.08 : . '

0.06 : .9. '
0.04 j.

0.02 -

CE Bladder Peak SD/Mean

 

 

0.00 4 L ‘ ‘ 7 ‘
0.000 0.001 0.002 0.003 0.004 0.005 0.006

CE Bladder Peak Height Mean

Figure 11. N on-effect of CE peak height on peak height variability.

45

0.0030 '
0.0025 ’
0.0020 "
0.0015 ’

0.0010 ’

abs/mg tissue

Oat

0.0005 '

0.0000 ‘ ‘ n * m H ‘ *

7.8 11.1 14.3 17.5 20.8 24.0 27.3
min

 

 

 

 

 

Figure 12. Comparisons of rabbit bladders metabolites from two different rabbit using
capillarly electrophoresis.

46

Injection Absorbance
T Duration Range

1 ‘-

10 s 0.10

 

20 s 0.20

JL

 

 

Figure 13. Variation in capillary electropherograms of rabbit bladder metabolites with
injection duration (and volume).

47

.a.
O

.0
co

 

.0
00

Relative Peak Height
0
K1

.0
o:

 

.0
01

 

-L

.0
.1;

 

A -J 1-1.ALIA1J L11]! 1 -l - '-1-l;11 '

3 4 5 101 2 3 4 5
Injection Duration (3)

.3
o I
O

[0

Figure 14. Change in selected rabbit bladder metabolite peak heights as a function of
injection duration.

48

 

 

’0'?
.§
0) '— . . . .
E 4‘5 18’ Prior Injection Duratlon
._ m _ .
5 0. 1 gsosec r=0.99
g g 1.7r
o. '43 ‘
Q) (D 1.6’
g 8 1.5:
0: m 14.
C (\l '
o _ .
-= s 1.3;
.2 c >
s 8 1.2;
o i
0 9. . .20360 r=0.88
(1;) g) 1.1’ Jlllaff'
7, g 1.0» tart-.9- ----- A nix-A2860 r=-0.11
Q) _
Q)
*‘m0.97“*“*“t““
V 8121620242832

min

Figure 15. Influence of prior capillary electrophoresis injection duration on a 2-sec test
injection of rabbit bladder metabolites.

49

more effected. These compounds tend to be acidic and therefore are attracted to the
positive end of the capillary. This end is also the sample injection end. During the run
the background buffer moves toward the negative end of the capillary. These acidic
compound show up later because, although they are carried by the background buffer,
they are retarded by the electric field acting on their negative charge. Injecting 0.1 N
NaOH returned the runs to their normal, higher speed. The acidic compounds appear to
remove Na+ from the capillary wall. The Na+ was replenished by the NaOH rinse.

During the protein experiments, it was necessary to rinse the capillary with NaOH before

every run.

Kinetics.

The assays for measuring Km's were the 3 ml assays done at 37 C. For each
substrate concentration three assays were simultaneously run. The substrate being
altered, either PEP or ADP, was prepared in solutions so that the volume added to the
assay was the same as with a normal activity assay, 100 111.

For the measurement of the activation energy the time course for the heating of
the thermal cuvette holder was first determined for the highest temperature. The SMPK
assays were continued until a change in the activity was observed while an assay was
running. The buffer was prepared separately for each assay using the temperature
dependence of the pH of an imidazole buffer and all the runs were done in one sitting to
minimize the effect of a change in observed activity due to a change in the PK solution.
The absorbance spectrum of SKPK at 30, 42, 47, and 71 C was used to determine the
temperature range for the Ea experiment. Figure 16 shows the spectrum at the first three
temperatures. When the assays with SKPK were done the activity did not show a

decrease in slope at 44 C. The SMPK assays were prepared for the range up to 75 C. The

50

 

 

 

 

 

 

0.5 '

0.4 ‘

1

<1) 03’
o
c
10
E

o 0.2“
U)
.o
<

0.1 *

0.0 '

 

260 280 300 320 340 360
wavelength (nm)

Figure 16. Temperature dependence of skeletal muscle pyruvate kinase absorbance.

51

substrates were all added without any enzymes and the cuvette was allowed to heat. The
final temperature measured during this heating step was the one used in the analysis.
Then LDH was added and a baseline activity was determined for that temperature. The
baselines were subtracted from the total activity measured. At the highest temperature
saturating pyruvate was added to determine the extent of LDH denaturation. The
absorbance had decreased to its final value before the spectrophotometer door was closed
showing that LDH was not limiting at this high temperature.

Tests with fructose 1,6-bisphosphate and phenylalanine were done using the 1 ml
assays at 25 C. The final concentrations were 300 uM for fructose 1,6-bisphosphate and

900 uM for phenylalanine.

CE coupling.

CE methods specific to the coupling experiments.

Norepinephrine (NE) and AA (ascorbic acid) ran widely separated in a borate
buffer at pH 7.4 and 9.4. A complex peak appeared at pH 9.4 and ran along with the NE
peak at pH 7.4 (figure 17). In the case of NE—AA binding at pH 7.4, both free NE and
the complex were in the same peak. A decrease in the observed NE peak could be seen
as the AA concentration was dropped. This was the technique later used for the
qualitative detection of binding in the kinases. At pH 9.4 the NE and complex peaks
were separated and reciprocal changes in the peaks were clearly seen as the AA
concentration was dropped. The ratio of the complex to the total NE present was
dependent on both the concentration of the AA and the electric field (figure 18). The data
points at each electric field were linearized (figure 19), and the parameters of the equation

were used to calculate the curves on figure 18. From the linearized plot a Ke could be

52

calculated (open triangle, figure 19). The Ke represents a dissociation constant of
molecules in an particular electric field. When the logarithm of the Ke's is plotted as a
function of the electric field a linear relation was seen and an extrapolation to zero
electric field gave a dissociation constant, (figure 20). In the case of the NE/AA the
dissociation constant was 74 uM AA when NE was 1 mM.

Using the linearization did not depend on using 1 mM NE as it also worked at 0.7
and 0.03 mM NE (figure 21). Using the points from figure 21, the dependence of NE-AA
binding on NE concentration was also confirmed (figure 22). This technique was used to
detect binding of NE to several other chemicals. Figure 23 shows the effect the
difference in binding of AA, AAP, and glucose. Other amines were also detected to bind
AA in a similar way (figure 24). Epinephrine had the best binding, Ke 7.9 mM, and
tyrosine the worst binding, Ke 27.4 mM, at an electric field of 179 V/cm. Serotonin did
not bind to NE. Morphine also bound to NE (figure 25) and a Ke at zero electric field
similar to that found for AA could be calculated (figure 26).

PK-CK CE interaction methods.

Angiotensin II antibodies were also shown to run the CE unit and so the technique
is appropriate for the investigation of binding in larger molecules. Separate peaks were
found for free antibody and antibody bound to Angiotensin H. The kinases were run
under conditions that were strictly limited because of the need to prevent adherence to the
capillary wall and the need to maintain EOF. For experiments on the coupling of
enzymes the buffer was 20 mM Na/PO4, pH 6.25, made up by mixing 20 mM NaH2PO4
and 20 mM Na2HPO4 in order to maintain a low ionic strength. Both CK and PK CE
peak heights had a strong pH dependence. PK's pH dependence was the strictest and
there was only a small peak at pH 6.3. The experiments were done at 6.3 because it is at

this pH that the absorbance effect was found to occur.

53

..,1

 

 

 

 

L04

pH 7.4

 

 

 

 

 

L0

LJLJ

 

 

 

17

L

CE Detection of Norepinephrine-AA Binding
1mM NE in 25 mM Borate, 204 V/cm

 

i

 

 

LN-1

50

40

 

00

pH 9.4

 

 

1111..

Ascorbic Acid (mM)

30 20

10

1

 

 

L__.1

Figure 17. Changes in capillary electropherograms of norepinephrine during ascorbic
acid additions in borate buffer at pH 7.4 and 9.4.

54

 

  
  
  
   

 

 

 

 

 

0 204V/cm
1_o: C1 179V/cm
8 09’ Z:158V/cm
1% 0'8 9 10:11::
II 0.7
g 0.6
g 0.5:
:1 31::
UJ
2 0.2
0.1 A—
0.0 .....
10'4 10'3 10'2 10‘1

Ascorbic'Acid (M)

Figure 18. Electric field dependence of norepinephrine-ascorbic acid binding.

55

 

 

log (Complex NE/Free NE)
0

 

 

_3....-44.LL4..
-2.7 -2.5 -2.3 -2.1 -1.9 -1.7

log (Ascorbic Acid (M))

Figure 19. Linearization of NE-AA binding and determination of the binding constant
and 95% confidence interval in CB experiments.

56

log Ke +/- 0.95 Confidence Value

log(Ke)

 

 

0 40 80 120 160 200
Electric Field (V/cm)

Figure 20. Extrapolation of The NE—AA binding constant to zero electric field from the
electric field dependence of NE—AA binding.

57

2 NE=10mM a.

 

 

 

-2.2 -2.1 -2.0 -1.9 -1.8 -1.7 -1.6
|0901:9014»

NE=07mM

 

 

 

-2.2 -2.1 -2.0 -1.9 -1.8 -1.7 -1.6
'09 (AA (M))

log (Complex NE/FreeNE)

 

 

 

-2.7 -2.5 -2.3 -2.1 -1.9
109(M(M))

Figure 21. Linearizations of NE-AA binding at different NE concentrations.

58

11

 

NE-AA Ke (M)

 

 

 

 

24 WW”...
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Norepinephrine (mM)

Figure 22. Concentration dependent change (fugacity) of NE-AA binding in an electric
field.

59

 

 

 

 

 

0.14 ' _._ NE
A 012: 'D'NE+100mMMP
g ' , -><- NE+5mMAA
-.-
g 0.10 _ NE+1OO mM Glucose
(D
g 0.08 '
g .
g 0.06 *
n .
<
UJ 0.04 '
z .
a .
a, 0.02
o_ .
0.00 ‘

 

 

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Norepinephrine (mM)

Figure 23. Effect of ascorbic acid - related compounds on CE binding to norepinephrine.

 

 

 

 

 

 

 

% 0.9 f ,0

E 0.8 , 9'

8 o 7 i

2 ' ; 30

2 0.6 _ .-

2 0.5 '

n .

8 0.4 '

o .

2 0.3 "

d) O 2 L -i- Epi-AA Ke = 7.9 mM

.E ° . ‘O" Tryp-AA K6 = 12.3 mM

< 0-1 T -I- NE-AA Ke=15.2 mM
0.0 ’ "O" Tyr-AA Ke = 27.4 mM

10'2 . 10"1
Ascorbic Acid (M)

Figure 24. Effect of norepinephrine - related compounds on CE binding to ascorbic acid.

61

log (Ke)

 

 

o 40 so 120 160 200
Electric Field (V/cm)

Figure 25 . Extrapolation of NE-rnorphine binding to zero electric field from electric filed
dependence of NE-morphine binding.

62

 

O NE-Ascorbic Acid
V NE-Morphine

 

 

 

log (Ke)

 

 

o 40 80 120 160 200
Electric Field (V/cm)

Figure 26. Comparison of NE-AA and NE-morphine extrapolations of CE determined
binding constants.

63

PK-CK interactions.

Sample preparations.

SMPK from the isolation and SKPK made up in an identical buffer were tested
against MMCK made up in a buffer that would bring the pH close to 6.3 when mixed in
know volumes. When required, the SMPK was filtered into a new buffer and lyophilized.
For the absorbance methods the SMPK was in diammonium citrate pH 5.3. When mixed
in a 10:9 ratio with CK in 20 mM Na/PO4, pH 6.7, the pH came to 6.25. SKPK was
again made up in the same solution as the SMPK.

For the absorbance measurements the enzymes were mixed at 25 C and the
solutions were allowed to develop for at least one hour. The development of the
solutions in a cuvette at 25 C was followed at 340 nm over an hour (figure 27). When the
solution was mixed with a plastic rod the solution only showed a slight drop in
absorbance. The absorbance did not rise much for the twenty minutes following this
point. A second mixing dropped the absorbance again to the same level. The choice of
340 nm was made in order to compare the experiments to previous ones and because the
greatest effect on absorbance was known to occur in two ranges, around 230 to 250 nm
and around 310 to 500 nm (figure 28). At the high wavelength end of the spectrum the
effect is more difficult to see because of the low absorbance of the enzymes.

The ethanol solubility experiments were conducted by adding an equal volume of
a water-ethanol solution to a solution of PK or PK+CK. The solutions were mixed and
allowed to stand for 10 minutes. They were then centrifuged at 10,000 g for twenty
minutes. The activity in the supernatant was then measured. This method of quantitating
the amount of PK in solution was chosen because its sensitivity required the least amount

of protein.

j

0.6

 

 

0.5”
00-4" f f
v
C? * mix mix
§o.3-
co
9
8
£02”
0.1-- - . . . . . . ..
0 10 2030 4050 6070 80 90
min“

Figure 27. Temporal development of absorbance change of combining MM creatine
kinase and skeletal muscle pyruvate kinase.

65

' 0.25
1.4r "'0— PK Abs

 

 

 

 

 

 

 

 

 

 

0 20 “"13— CK Abs
1'2, ' + PK-CK Abs
1.0 '
g 0.8' g
E e
g 0.6“ 8 0'10
< 04’ 3
' 0.05
0.2
0.0 0.00 ’
220 260 300 340 400 600 800 1000
wavelength (nm) wavelength (nm)

Figure 28. Spectrums of skeletal PK alone, MMCK alone, and a 0.5-0.5 mix of SKPK
and MMCK.

66

The activity measurements were done with several PK and CK concentrations
with 0% ethanol. The lowest PK concentration in the original solution was 144 ug/ml
while the highest was 2 mg/ml. It was thus discovered that CK had an effect on PK
activity that did not require the low Mg++ assays previously used (Dillon, et. al, 1995).
These assays were repeated in two ways. First the procedure was repeated with neither a
centrifugation nor a development step. Second, an assay was started with PK alone and

CK was added halfway through the assay.

67

RESULTS

Isolation of smooth muscle pyruvate kinase.

PEG (polyethyleneglycol) with an average molecular weight of 3,350 was able to
precipitate PK to the exclusion of other proteins. The brought up precipitate showed little
loss of activity with a six to eight fold increase in specific activity.

Each step in the second procedure chromatography yielded better than a 5 fold
increase in specific activity except for the gel-filtration and the second affinity step. Both
of these were still required. The gel filtration was required to prepare the post-PEG
sample for the CM chromatography and the blue gel because no other step yielded an
increase specific activity at that point in the isolation. From 200 g of bladder, the
urothelium was removed and not weighed during this isolation. At the start there was
82.5 g of protein and 6430 U of activity. The fold purification during the best two steps
was 7.52 for the PEG step and 5.72 for the CM chromatography. The gel filtration gave
the poorest recovery: 48%. The final specific activity was 294 U/mg at 37 C and a total
of 1400 U were collected. The last chromatography step was run in 5 steps and the eluent
collected had a concentration of 0.160 mg/ml before lyophilization. The SDS—PAGE
(figure 29) revealed that after the CM step there was one major protein contaminant with
a high molecular weight. At this stage the isolated PK band was still very faint when
only coomassie staining was used. The SDS-PAGE showed that when 1.25 mm thick
gel was overloaded the sample did contain two minor bands other than PK. These bands

only showed up when the gel was silver stained. SKPK from Sigma Chemical Co. also

68

 

 

 

 

 

 

 

 

 

 

 

 

 

 

J
I:
J
l

J
J

 

 

 

 

 

PK PK CK

 

 

 

 

CM step lst affinity step 2nd affinity step skeletal enzymes

Figure 29. SDS gels of smooth muscle and skeletal muscle pyruvate kinase.

69

showed extra bands. The isolation of SMPK produced purification similar to that of
commercial SKPK.

Each bulk step in the first and second procedures was tested on the homogenate
supernatant and every time the CK activity was separated from the PK activity. It was not
possible to co-isolate the proteins as a complex. The final isolated PK showed no CK
activity.

The first isolation technique was also tried using porcine tissue up to but not
including the affinity chromatography. The results were similar in every case except that
the ammonium sulfate gave an increase in specific activity only when it was used as the
first step in the porcine isolation. In the bovine isolation it was not successful.

Bladder smooth muscle metabolites run on the CE made it possible to assess the
effect of the period spent on ice (figure 30). The peaks that were identified by using the
technique developed with the rabbit bladders are numbered. In figure 31 the level of the
different metabolites is given for five samples taken from five different bladders.
Although the PCr was low and the ADP was high, there were still large triphosphate

peaks and the lactate level remained low.

Kinetic properties of skeletal and smooth muscle pyruvate kinases.

The Lineweaver-Burk plots for ADP were linear as expected (figures 32 and 33)
and the Km's were similar: 254 uM for SKPK and 214 uM for SMPK. When plotted on
the same graph by using the specific activity they appeared similar (figure 34). Both
samples were lyophilized and gave low specific activities calculated from the Vmax.
Figures 35, 36, and 37 show the same graphs for PEP. The Km's are 40 uM for SKPK
and 86 uM for SMPK. When phenylalanie, ATP, and fructose 1,6-bisphosphate were

70

,f

12

WWM'~

diSll
diVOl

1

AA L...

 

dOVG

 

13d?

7

5
-flLMWWA

arena 9

2 min

FVW WW.)

 

 

V3000

(

Figure 30. Capillary electropherogram of beef bladder extract.

71

 

 

 

 

 

 

 

 

 

 

‘5
0.014 - <
. O)
0.012 r I
0.010 -
8 0.008- a.
i ‘ 4
§ 0006’ go
i 3 a: 5 :
0.004 - 3g 0. II
I o co
* .0
0.002 - 1 5‘° 7 ll
1 2 4 12
0.000 n n n. .llll ['1 . . n
10.3 14.4 13.5 22.6 26.7 30.8 34.9

min

Figure 31. CE-measured beef bladder metabolites from five bladders used for smooth
muscle PK isolation.

72

60 “

 

 

-10 -5 0 5 10 15 20
1/ADP (mM)

Figure 32. Lineweaver-Burk plots of ADP dependence of skeletal muscle PK.

73

500"

400"

 

300"

min

200‘”

 

1001-
5 0 - 5 10 15 20
1/AD~P(mM)

 

Figure 33. Lineweaver-Burk plots of ADP dependence of smooth muscle PK.

74

80’

 

70’ I skeletal
.. O smooth

 

 

 

 

min x microgram

 

 

-5 0 5 10 15 20
1_/ADP (mM)

Figure 34. Comparison of the ADP-dependent kinetics of SKPK and SMPK.

75

50"

 

 

 

10 20 30 40 50
1/PEP (mM)

 

Figure 35. Lineweaver—Burk plots of PEP dependence of skeletal muscle PK.

76

500 ’

400

300 '

min

 

 

 

100
-20 -10 0 10 20 30 40 50

1/PEP (mM)

Figure 36. Lineweaver-Burk plots of PEP dependence of smooth muscle PK.

77

 

70’

   
     

I skeletal
60 O smooth

 

 

 

50'

40”

30'

min x microgram

20'

 

 

-30 -20 -10 0 10 20 30 40 50
1/PEP (mM)

Figure 37. Comparison of the PEP-dependent kinetics of SKPK and SMPK.

78

tested as effectors of peak SMPK activity no effect was found. The Vmax values in U/ug
determined using ADP (0.071 for SKPK, 0.042 for SMPK) were similar to those
determined using PEP (0.064 of SKPK, 0.040 for SMPK).

The temperature dependence of the activity of SMPK and SKPK is shown in
figure 38. For SMPK a decrease in the expected activity was first seen at 55 C but the
slopes of the individual assays did not appear non-linear until 60 C was reached. The
Arrhenius plots (figure 39) of both kinases were similar with neither one showing a
prominent bend at the non-denaturing temperatures. The activation energies calculated
from the least squares regression line were 56,400 for SKPK and 61,100 J/mole for
SMPK.

CE coupling experiments.

The absorbance relation at 195 nm produced a decrease in absorbance when
SKPK was added to MMCK, as opposed to the increased absorbance seen at 340 nm
(figure 40). The SKPK was strongly effected by the MMCK in concentrations greater
than 1.0 mg/ml. The SMPK absorbance was not strongly effected by the presence of
MMCK. When CE is used on MMCK the peak area is not linearly related to the
concentration of CK (figure 41) as it is for metabolites. Yet an effect similar to the NE-
AA effect can still be seen when MMCK is run in combination with SKPK (figure 42).
As expected from the measurements of absorbance at 195 without using CE (figure 40),
the combined MMCK/SKPK peak was lower than expected from the individual
absorbances. When this experiment was done using a constant SKPK concentration, an
effect on the combined peak could be seen (figure 43). The ratio of the peaks including

the complexes and the peaks of CK run alone did not approach 1.0 (as it would if there

79

190'
170’
150'
130'
110'

90“

70'

PK Activity (U/mg)

50'
30'
10’

 

10

Figure 38. Temperature dependence of SKPK and SMPK activity.

 

 

+ Skeletal PK
“0“ Smooth PK

 

 

L

20 30 40
Temperature (C)

80

 

50

60

91.01
ON
101

  
 

 

(specific Vmax)
.00 P P P P P
(D O N -b O) 00

9 Skeletal PK
0 Smooth PK

 

 

 

n
NAG)

9°
0

 

3.0 3.1 3.2 3.3 3.4 3.5
1000/r (K)

Figure 39. Activation energy determination of SKPK and SMPK.

81

  
 

 

 

 

 

 

1.1 '
1.0 ~ .240
1‘ 0"
0.9 '
E 0.8 -
8 .
1: 0.7 ’
a, L
8 0.6‘
g .
8 0.5:
g 0 4 _ '0" ExpectedAotCK+SKPK
. + MeasuredAofCK+SKPK
0.3 'A“ ExpectAofCK+SMPK
0 2 _ + MeasuredA of CK+SMPK
L

 

0.0 0.4 0.8 1.2 1.6 2.0
Creatine Kinase (mg/ml)

Figure 40. 195 nm measurement of the expected absorbance of solutions containing both
PK and CK from the absorbance of individual solutions containing SKPK, SMPK or
MMCK.

82

0.04 '

0.03 '

0.02 ‘

0.01 ’

CE Peak Area (Abs-min at 195 nm)

 

0.00 '

 

0.0 0.4 0.8 1.2 1.6 2.0
MMCK (mg/ml)

Figure 41. Concentration dependence of the peak areas of capillary electropherograms of
solutions of MM creatine kinase.

83

 
   

. 1 x02:
8k 0 1 Eva + v6.2.2

 

x35 :59: 4.0 + 52.2 .535 m

 

viva .565 to m

will.

58 m

‘ < mood

v6.2.2 .536 N

Figure 42. Capillary electropherograms of MMCK, SKPK, and their combination.

84

1.2 '

1.1”

1.0 ‘

0.9 '

0.8 ‘

(MMCK + SKPK)/(MMCK) - 195 nm

 

 

 

0.7 . . ....... . . . . ..
0.0 0.4 0.8 1.2 1.6 2.0

MMCK (mg/ml)

Figure 43. Ratio of the CE absorbance of MMCK + SKPK to the absorbance of MMCK
as the creatine kinase concentration increases at 204 V/cm.

85

 

 

 

 

 

 

0.09 r
E 0.08 -
c 0 07 L
8 ° ’ ‘0' 204 V/cm
5 0.06 "
a, .
8 0.05 -
9 .
a: 0.04 '
E .
D 0.03 '
8 .
c 0.02 -
m .
e 0.01 ~
0 .
8 0.00 -
< ,

-0.01 ‘

10'1 10°
MMCK (mg/ml)

Figure 44. The CE absorbance difference between MMCK and MMCK + SKPK as the
MMCK concentration increases at 204 V/cm.

86

were no interactions) as the MMCK concentration was increased. Figure 44 shows the
difference between the MMCK peak and the combined peak. As the MMCK

concentration increases, the difference between the peaks gets larger.

PK-CK interactions.

The absorbance of PK-CK interaction was measured at 340 nm by matching the
most prominent species in the two tissues (figure 45). MMCK and SKPK showed a
marked increase in absorbance above that expected for the addition of the two solutions.
The BBCK and SMPK combination did not show this relationship. When the effect of
BBCK was tested on SMPK solubility and activity in ethanol it was found to have no
effect (figure 46). The same relationship was shown when testing SMPK versus MMCK
(not shown). When the same experiment was done with SKPK and MMCK a new effect
was discovered. A marked increase in activity of SKPK was seen and it persisted up to at
least 26 % ethanol (figure 47). Without ethanol there is a larger relative effect at lower
SKPK concentrations than at higher SKPK concentrations (figure 48). The SMPK, at a
concentration between the two SKPK measurements, was not effected by either BBCK or
MMCK. But the effect was seen when SKPK was combined with BBCK. When the data
are plotted using specific activities (figure 49), the SKPK concentrations both higher and

lower than those used for the SMPK tests still show a marked effect of MMCK.

87

0.25 ’

0.20 '

0-15 * MMCK + SKPK

0.10 '

0.05 ' ¥

0-00 ' BBCK + SMPK

-0.05 - KB

0.0 0.4 0.8 1.2 1.6 2.0
CK(mg/ml)

Extra Absorbance (340 nm)

 

Figure 45. Comparison of the 340 nm absorbance differences produced by combinations
of MMCK + SKPK and BBCK + SMPK.

88

8

 

 

 

  

 

 

 

 

 

 

 

 

 

w
m. a 1%
hwnb
a +
Kuun .
p p 1%
M M . \
euna .
V. mmaflammfiflmflflflmmw
s E
\ \ \ k \A
’1 P l, i. y, x ( l .l, .157 ’1} ’1’1 ’1.

 

 

7 6 5 4 3 2 1 0

33:: 23:03. E 2822 586m

0

17 20 23 26 29 32

°/o Ethanol

Figure 46. Effect of BBCK and ethanol on SMPK activity.

89

 

E3 SKPK alone
5 SKPK + MMCK

 

 

 

Skeletal Muscle PK Activity (mUa/ul)

 

O 17 20 23 26 29 32
°/o Ethanol

Figure 47. Effect of MMCK and ethanol on SKPK activity.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

22.3.5 23:2 vi

-SKPK-
+BBCK

-SMPK- -SMPK-
+BBCK +MMCK

+MMCK

------—SKPK-------
+MMCK

Figure 48. Effect of both MMCK and BBCK on the activity of SKPK and SMPK.

91

94040404 «0404 4 4.. 404.404.404.494.40404040404304
.. s Kane».woweamuvvvvvwvvveg.» we.» . e

 

 

 

 

 

   

 

 

 

 

     
   

 

 

 

 

q 4.40404 404010104040«0104040401040101
.. §e$vvvvvvvfislfi£3.»
8 6 4 2 0 8 6 4 2 0
1 1.. 1 1 1
855v 222 Va

-SKPK-
+BBCK

-SMPK- -SMPK-
+BBCK +MMCK

+MMCK

-------SKPK-------
+MMCK

Figure 49. Effect of both MMCK and BBCK on the activity of SKPK and SMPK per unit
92

of PK.

DISCUSSION

Smooth muscle PK isolation.

The isolation is unusual in that one of the most successful isolation steps was one
of the bulk methods. The final PK isolated had a specific activity close to that of SKPK.
Also, the SKPK had a few minor contaminants when run on SDS-PAGE. Considering
that the effects of CK on SKPK were clear, it is unlikely that the lack of observed effect
with SMPK was due to interference by contaminants. For this to be the case the
contaminant would need to have a specific effect. The failure of a CK-PK co-isolation
was not surprising since the coupling is a transient one. It was even less surprising when
it was discovered that CK did not have the effects on SMPK as it did on SKPK.

There was a possibility of tissue degradation when the bladders were kept on ice.
This was especially true since the entire batch of bladders had to be put through the first
step of the isolation as soon as the tissue was brought to the laboratory. The capillary
electropherograms showed that the tissue was still in good shape after the time spent on
ice. The high tri-phosphates especially showed that the tissue had not used up its energy
reserves. The low lactate levels reflect low use of glycolysis. Most of the energy used
was derived from the stored PCr pool. The ability to keep the tissue on ice for an
extended period of time was useful since freezing in liquid nitrogen caused some loss of
activity. Also of importance was the fact that no broad raised area appeared in the
electropherograms; figure 30 shows a level baseline. Such a broad peak would have

been a sign of proteolysis as seen in previous experiments using trypsin.

93

Properties of SKPK and SMPK compared.

The two enzymes appear similar except that the SMPK has a lower specific
activity. As seen in figure 48, where the activities of the two SKPK experiments
bracketed the activity used in the SMPK experiment, this lower specific activity of the
isolation sample cannot explain the difference in the activity effect of CK on the SKPK as
opposed to SMPK. All the other parameters were similar for SMPK and SKPK. The use
of lyophilized sample made the SKPK specific activity calculated from the Vmax low.
The same relationship was seen in the Vmax calculated from the Lineweaver-Burk plots
as from the specific activity of SKPK and SMPK measured with saturating conditions
during assays at 37 C. The SMPK Km for ADP of 214 uM is high compared with the
resting ADP in smooth muscle, which has been measured at 30 uM (Fisher and Dillon,
1988 ). Under severe conditions the ADP has been shown to reach 120 uM (Fisher and
Dillon, 1988 ). Increases in ADP therefore should significantly increase flux through
SMPK. As mentioned in the background though, changes in glycolytic flux have been
observed which can not be explained by changes in the substrate concentrations for PK.
In such cases, the increase in PK activity is assumed to come from another effect. There
is then a dissociation between the effects of changes in substrate concentration in vitro
and the effects of changes in substrate concentration in viva. Channeling of metabolites
has been suggested as an explanation for these mismatches but the lack of interactions
between SMPK and CK suggests that this does not involve CK. The lack of effect of
phenylalanine, ATP, and fructose 1,6-bisphosphate on the activity of SMPK under
saturating substrate conditions suggests that any effect of these could only be on the Km
of ADP or PEP. The effect on the SMPK Km would not be comparable to an effect in
SKPK because of the transfer of metabolites between SKPK and MMCK in skeletal

tissue.

94

The Activation energies for the two enzymes are similar at 56.4 and 61.1 kJ/mole.
They are also comparable to other enzymes. It is interesting that there was a large change
in the absorbance spectrum of PK (figure 16) when it was heated to temperatures at which
the activity measurements did not show any denaturation (figure 38). The absorbance
measurements were all done using a single buffer so there may be a pH effect due to the

changes in temperature.

CE coupling experiments.

The original idea for measuring weak interactions was to measure the diffusion
coefficients of molecules in capillary tubes in the presence or absence of other molecules.
The new technique developed is more efficient for several reasons. A model for the
change in shape of the peak does not have to be produced since a empirical linearization
was possible. The experiments can be run as normal CE runs instead of allowing the
molecule to diffuse overnight. And most importantly a new parameter can be measured,
the Ke. This dissociation constant measured in a particular electric field is not just used
to extrapolate the dissociation constant in zero electric field. It can be used to estimate
the effect of naturally occurring electric fields on biomolecules. The electric field
generated by a membrane's surface charges can be calculated as a function of the distance
from the membrane in an aqueous solution (figure 50). Using the Ke's determined for
NE-AA dissociation and the distance dependence of the membrane's electric field, it is
possible to plot the Ke as a function of the distance from the membrane (figure 51).
Putting both graphs together (figure 52) it can be seen that the NE—AA complex is
expected to dissociate completely before it reaches the membrane. The experiments
using NE and AA led to the theory that, for a hormone that binds to another molecule, the

concentration of free hormone increases close to their binding sites on membranes

95

 

 

2.0E5 :
1.855 j
E 1655 f E = dV/dx = (-1/ MVO exp(-x/?t)
:5, 1.4E5 :
E 1.2E5 ’
.9 '
LL 1.0E5 ’
8 ;
.3 8.0E4 »
m 6.0E4 :
4.054 :
2.0E4 j
0.0E0 ’

 

 

0 2 4 6 8 10 12
nanometers

Figure 50. Estimation of the electric field generated by a cell membrane.

96

0.04 ’

 

 

-0.01 . . . - . . - .
7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8.0

Distance from Membrane (nm)

Figure 51. Estimation of the change in NE-AA dissociation constant at different
d‘5‘ances from a cell membrane.

97

 

 
   
 
 
 
   
 

 

 

 

 

 

2055 0.4 nm complex
1.855 x 0.04 .
1.6E5 § 0 03 _
A 1.455 E
E E g L
§ 1.255 -E v 0-02
V a): £2
2 5‘.
.9 1.0E5 _|E 001..
u. '0:
.g 8.054 g;
‘6 2735 0.00 " """"""""""""""""""
£60545;
E 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8.0
4.0E4 5 ~\ nm /’
3 \\ //
2.0E4 E \\ //
i \\ //
0.050 '
0 2 4 6 8 10 12

Distance from membrane (nm)

Figure 52. Experimental range of Ke measurements of NE-AA dissociation relative to
the electric field generated by a cell membrane.

98

because of the electric field generated by the surface groups on the membrane. This
method was applied to SKPK-MMCK binding and gave qualitative evidence that the
binding occurs. This suggests that the same effects may be seen on the inside of
membranes (where a smaller but similar electric field exists) with interacting proteins as
on the outside of membranes with hormones.

The failure to show, with the PK-CK interaction, a quantitative relationship
similar to that shown for the smaller molecules was due to several factors. First the
conditions under which PK and CK run through a capillary are restricted by the need to
prevent adherence to the capillary walls. Second, the concentration dependence of the CE
peak area, as measured using CK, is non-linear. This may also be due to adherence. The
size of the molecules is not in itself a problem because the angiotensin II antibodies

showed two peaks, a free and complexed peak.

Specific properties of pyruvate kinase and odd characteristics.

Problems in the current understanding of PK.

PK has several odd characteristics. The mass-action ratio for PK suggests
that its reaction is far from equilibrium while its maximal catalytic activity suggests
otherwise (Newsholme and Start, 1981). It is thought that mammalian muscle PK is
usually inhibited because of its exceedingly high activity (Newsholme and Start,
1981). In vivo ATP and the low concentration of PEP compared to PK's Km may
cause this inhibition (Newsholme and Start, 1981). Yet the concentrations of these
substrates do not change enough in cells to make a noticeable difference in PK
activity. In a case where glycolytic flux is increased 100-fold there was no observed

change in PEP concentration (Sacktor and Wormser-Shavit, 1966). The increase in

99

activity in this system can not be the result of decreased ATP or increased PEP

concentration.

The diazymatic PK and CK reaction.

In previous research PK and CK were shown to bind and to exchange ATP
formed from their respective reactions without ATP being released into the bulk
medium; this was named a diazymatic reaction (Dillon and Clark, 1990; Dillon et al.,
1995). During NMR experiments, Hseih and Balaban (1988) may have inadvertently
measured this ATP exchange in vivo when they saturated in the position where PEP
would be. Since the transit time of the phosphate at the position where saturation is
tacking place is always going to be longer than the time it takes to damp it, it is the
flux of the metabolite through a point being saturated which is important in
transferring a damping signal to another metabolite. Although PEP is not seen, it is
still being damped and if the flux is high through the diazymatic reaction the damped
signal will accumulate in the phosphocreatine pool. When Balaban saturated the
position where PEP is located, there was a decrease in the ratio of the
phosphocreatine to yATP peak compared to the experiment during which there was
no PEP saturation.

Only a few facts about the mechanism of the diazymatic reaction can be
deduced: the rate of exchange was measured with the NMR; the PK reaction, which
is associative with respect to PEP hydrolysis , precludes the possibility of directly
transferring the phosphate from PEP to creatine bound on CK; and it is also not
possible to isolate phosphate exchange between PK and CK without ATP in the

solution since there is no other intermediate for the PEP phosphate.

100

Creatine kinase and channeling.

Specifics of CK and channeling.

CK is better understood than PK. One of the most interesting areas of
research in CK is the phosphoryl-creatine circuit. An original creatine phosphate
shuttle hypothesis was proposed by Bessman and Geiger in 1981. The phosphoryl-
creatine circuit is a modified version of the phosphocreatine-shuttle hypothesis of
Wallimann and Eppenberger (1983). Antibodies against M-CK have been used to
localize CK in chicken skeletal muscle. This enzyme thought to be free in solution
was actually found on the M-line and I-band but not on the Z-line. The
solubilization of the plasma membrane and washing of the tissue prior to fixation
caused the I-band CK to disappear but had no effect on the M—band (W allimann et
al., 1989) This M-band bound CK amounted to 3 to 7% of the total CK. In general
the phosphocreatine circuit postulates that the sources of ATP generation are
connected to ATP usage sites via phosphocreatine. High energy phosphates from
the mitochondria are translocated to the cytoplasm by Mi-CK in the form of
phosphocreatine. This is supported by the close association of Mi-CK with the
ADP-ATP translocase at points of contact between the inner and outer mitochondrial
membranes. The hypothesis also supposes glycolysis as an entry point for ATP into

the circuit. If this is true, diazymatic interaction of PK and CK may be involved.

Significance of CK's properties.

The functions of the circuit which have been proposed are energy buffering,
energy transport and channeling, and regulation of local ATP levels. One interesting

observation is that B-CK has two isoforms and may be phosphorylated as

101

demonstrated with two-dimensional gel electrophoresis (W alliman er al., 1989, p.
168). This fits well with the concept that isoforms of an enzyme may have different
affinities for other enzymes or subcellular structures and thus may differentially
distribute their kinetic properties to various pathways (Ureta, 1991).

Although many tissues work well without CK, its importance is probably
greater at high workloads. There is considerable evidence for this and in many of
these cases CK is also associated with subcellular structures. Training for long
distance running and the resultant adaptation of fast-twitch to slow-twitch fibers
occurs at the same time as a decrease in total muscle CK and a increase in Mi-CK
(W allimann et al., 1989, p. 171). This was used to support the concept of a dual
function for the phosphoryl-creatine circuit as an energy buffering system in acute
energy needs and as an energy channeling system in prolonged energy needs. There
has also been positive correlation between oxygen consumption and CK reaction rate

by 13-P NMR, biochemical measurements, and model calculations (W allimann et

aL,1989,p.171)

Smooth and skeletal muscle PK-CK interactions

The absorbance measurements at 340 and 195 nm both showed an effect of CK on
SKPK. The MMCK is the CK isoform found in skeletal muscle. Smooth muscle
contains mostly BBCK but it also contains a small amount of MMCK. The fact that there
was no effect of either BBCK, tested at 340 nm, or MMCK, tested at 195 nm, on SMPK
was the first strong evidence that SMPK acted differently from SKPK. The
measurements at 195 nm were only possible by injecting the samples into the CE unit.
The spectrum measured down to 220 nm using the spectrophotometer always showed a
positive change in absorbance upon interaction of the proteins. This change was not as

great at 220 nm but was never negative. The negative effect seen at 195 nm also showed

102

a difference between the SMPK and the SKPK. This effect at 195nm in combination
with the changes in the strength of the positive effect seen at higher wavelengths suggests
that there may be some specificity to this interaction. In the spectrum measurements on
figure 28 the relative difference between the CK and PK curves decreases as they
approach 220 nm. If the absorbance of the interacting species continues to changes in the
same direction as the wavelength decreases, this could cause two coupled molecules to
have less absorbance than the free molecules. The steepness of the curves at this position
makes it impossible to accurately compare them from the two or three points on the rising
portion of each curve.

The ethanol solubility experiments worked well for SMPK and showed no effect
of either MMCK or BBCK. Since no effect of CK on SMPK was seen at 0% ethanol it is
assumed that the SMPK activity accurately measured the amount of SMPK in the
solution. It is interesting to note that earlier ethanol experiments using SDS-PAGE to
quantify the amount of SMPK remaining in solution needed more ethanol to decrease the
level of PK in solution. This may be due to differences in protein concentration or in
buffer, or it may reflect an effect of ethanol to decrease the activity of PK. It is also
interesting to note that the SKPK required higher ethanol to start to decrease in activity
(figure 47) than the SMPK (figure 46). The effects seen in the SKPK experiment are
complex. There are two effects which cannot be separated. First there is the effect of
MMCK to increase the activity of SKPK. Since the SKPK+MMCK 23% sample (figure
47) has as more activity than the SKPK 0% sample, the activity effect of MMCK must
still be occurring at this level of ethanol. At greater than 23% ethanol, when SKPK starts
to decline, it is not possible to determine what fraction of the increase in activity is due to
a direct activity effect as opposed to a solubility effect.

BBCK and SKPK are not found in the same tissue. Despite this, the strong effect
of BBCK on SKPK activity suggests an important role for the type of PK found in a

tissue. It may be that it is the PK isozyme found in a tissue which determines whether it

103

will couple to CK. In this case the property that makes it possible for CK to couple with
SKPK is found on all CK isozymes. This also suggests that there are surface structures
on CK isozymes which are identical in all the isozymes and which are complementary to
structures on the SKPK isozyme but not complementary to structures on the SMPK
isozyme.

The energy fluxes in skeletal muscle can be characterized as changing rapidly and
to extreme values compared to other tissues. Smooth muscle has many of the same
characteristics as skeletal muscle because of their similar function but smooth muscle
never experiences the changes in energy demand occurring in skeletal muscle. The
possibility that a channeling mechanism between the energy producing pathway (PK) and
energy buffering pathway (CK) exists in skeletal muscle but not in smooth muscle is
significant. It suggests that, in the case of the PK-CK interaction, the effects emphasized
by the modern view of metabolism are part of the design of the metabolic pathway that
makes it specific to the needs of the tissue. Because of the much greater glycolytic flux
in skeletal muscle than in smooth muscle, PK-CK coupling may play a central role in

skeletal muscle energy metabolism.

A general modern view of the cell.

When thinking of the cell as a gel of continually interacting proteins and of a
number of these interactions leading to functional coupling, is there a problem with
aggregation due to weak interactions? Even if the interactions are weak, if the
concentration of molecules is high enough, one would expect that the number of
interactions would lead to a substantial bound fraction at all times. This would tend
to cause aggregation which implies precipitation. It doesn't seem possible that, even

if two proteins help solvate each other, a large number of proteins interacting in this

104

way would cause a concomitant increase in solvation. On the contrary, one would
assume that an optimum is quickly reached and that further interactions would be
counterproductive. In cells, the largest multi-enzyme complexes in solution are
ribosomes and even their "soluble" population may be bound to the sub-cellular
lattice. Past a certain molecular weight molecules are difficult to keep in solution.

In the case of glycolytic enzymes such as PK, large multi-enzyme complexes
have been suggested (Kurganov et al., 1985) because of the number of interactions
found between many of theses enzymes. These structures have not been seen in
electron rnicrographs of muscle tissue where they were proposed. The interactions
observed may therefore be due to sites which act transiently. The problem of
precipitation would be especially important in complexes which did not act
transiently if the proteins were globular.

The problem of precipitation is avoided in structures such as the sub-cellular
lattice and the contractile apparatus by forming molecules which have a long axis
and whose interactions define a minimum distance. Research has found that, in cells
which need to form a gel phase for the purpose of locomotion, the gel phase is
created by long actin filaments bound together by an actin binding protein which has
angled binding sites so that the filaments must cross at ninety degree angles in order
to bind to each other (Stossel, 1994). As viewed in electron micrographs, there is an
orthogonal network of strands. In other words, the cell has maintained a minimum
distance between the gel forming molecules so that there is solution space between
them. Although these molecules have, in effect, precipitated, they have done it in a
specific way which has spared other proteins.

It is known that water molecules, sometimes large networks of water
molecules, are trapped in locations within folded proteins and isolated from the bulk
solvent (Creighton, 1984, p. 243). Similar networks may form as proteins approach

and their highly ordered hydration spheres become interrneshed. Just as in the inside

105

of proteins, these water molecules may be an integral part of the ionic interactions.
They may be important during weak interactions to keep the proteins from
aggregating. Also the concept of complimentary is useful in that, if many of these
interactions are through specific sites but that there is a large number of types of sites
and a limited number of sites per molecule, then some molecules will be
complimentary only to certain other molecules. Once these molecules have bound,
the sites on these molecules will be taken up and they will not bind further and
aggregate.

The number of weak non-specific interactions between molecules and the
concentration of these molecules may be optimized quantities. Different cell types
may have different optima and different modes of operation of individual cells may
change the optimum, leading to the regulation of these parameters. For example, a
macrophage which starts to migrate in response to a chemotactic stimulus may need
to change these parameters in order to facilitate cytoplasmic streaming. Even local
areas within cells are apt to maintain these parameters differentially.

It is also interesting that many of the enzymes thought to associate for
functional reasons tend to associate with sub-cellular structures. This may be a way
of forcing these dynamic interactions to occur on supports which cause them to be
separated in space. This would also mean that the need for cellular support space

may be an driving evolutionary principle.

The relativity of specific weak interactions.

Is there a problem with specific weak interactions as opposed to non-specific
interactions? Non-specific interactions are weak. If an interaction is strong then it

will be specific. The words weak and non-specific are almost defined in terms of

106

each other. The term weak is relative and intermediate between non-specific and
strong. Although experimentally these interactions require high protein
concentrations and low ionic strengths, these are ways to bring the solutions closer to
the cellular conditions (Brooks and Storey, 1991). In the case of PK and CK low
ionic strength is not a necessity.

It is still possible for weak interactions to lead to specific functional coupling.
Functional coupling will only occur through the exchange of metabolites which are
specific to the binding sites of the enzymes and allows the proteins to come together
with the alignment which will optimize transfer. It is easy to see that, when non-
specific interactions between two defined surfaces--one on each protein-~occur, the
degrees of freedom will be restricted by the approaching electrostatic charges.
Electrostatic interactions have the potential for acting at greater distances than other
types of interactions. The other interactions; hydrophobic, Van der Walls, and steric;
in combination with these electrostatic interactions also have all the effect of
reducing the degrees of freedom of motion of the proteins when they are adjacent.
These may not be strong enough to cause a strong interaction, but if the proteins
come in contact they may be forced into a relative orientation which favors a
functional interaction. This is the essence of non-specific interactions. It is not
possible for weak interactions to bring proteins together in dilute solutions and force
functional coupling to occur. But in more concentrated solutions where interactions
will occur, the weak interactions can cause the correct orientation of the proteins
when their appropriate surfaces collide. A protein with a large dipole moment, like
cytochrome c, could use these interactions from further away. Such dipoles caused
by a—helices have been shown to aid in the binding of substrates (Hol et al., 1978).
Although the substrates which bind are generally in hydrophobic clefts which help to

shield them from a solvent reaction field, protein interactions have the advantage that

107

both species can develop these large dipole moments, and allosteric interactions may
help release metabolites.

It may not be necessary for enzymes to be tightly bound for their substrates to
be exchanged. It may simply be necessary for enzymes to be close enough to each
other compared to other enzymes for the diffusional distance of metabolites to be
reduced and for channeling to take place. This possibility has been used to explain
the increase in activity seen in sequential enzymes which Mosbach had bound to a
solid support(Ottaway and Mowbray, 1976, p. 141), Ottaway and Mowbray have
modeled this possibility and found it to be plausible(0ttaway and Mowbray, 1976),
and "Bernhard and Srivastava ...... argued that 'relative thermodynamic stability of
protein-protein complexes has no bearing on the channeling phenomena"'(p. 193,
Ovadi, 1991).

The mechanism by which complementarity yields specificity in weak
interactions may also be important. This is especially true if these transient
interactions have allosteric effects which increase the rate of product release. Both

steady state fluxes and channeling may be increased by this mechanism.

Current protein interactions research.

Most of the current research in protein interactions does not involve enzymes
which act in a hand-off manner. The research on interactions between subunits of
proteins is extensive. Interactions between separate molecules are also being
investigated. For example the interactions between endophilin and amphiphysin
(Micheva et al., 1997) are being investigated in a manner than emphasizes the
involvment of particular protein domains. New techniques include methods of

studying protein interactions which can also be applied to enzymes that hand off

108

substrates. Such techniques include "high-throughput screening strategies" (Kay and
Paul, 1996).

The increased speed of desk-top computers has resulted in an increase in
modeling. The most extensively used method is known as Brownian Dynamics
(Northrup, 1998). In this approach, two proteins of know structure are used. One
protein is assigned a central position surrounded by a sphere whose diameter is
assigned by the modeler. The second protein is placed at an initial position on the
sphere. The algorithm takes the structures of the proteins and the electrical fields
that they generate in determining the movements of the second protein. When the
program is started, the second protein is allowed to take a random walk. Several
criteria can be set to test if the proteins have bound. The simplest test is the setting
of an arbitrary distance between the two proteins. The program stops when either of
two conditions is met--the two proteins have bound (as defined by the programmer),
or the second protein has left the sphere. By running the program repeatedly, a
probability of docking can be determined and compared to that of other proteins.

Also of interest is the relative orientation of the two proteins at the time of docking.

109

SUMMARY

The interactions of SKPK and SMPK with MMCK and BBCK were investigated.
An isolation procedure was developed for SMPK and it was found to have properties
similar to SKPK.

In the course of the investigations on interacting molecules, a CB technique was
developed to measure binding between molecules. This technique made it possible to
quantitatively measure the effect of the electric field on the dissociation of molecules that
bind. This technique can be used to estimate the effect of the electric field of
biomembranes on molecules that bind to each other.

It was determined that SMPK does not interact with CK isozymes found in
smooth muscle the way SKPK interacts with MMCK in skeletal muscle. It was also
found that it may be the PK isozyme that determines whether the interaction with CK will
occur.

The modern view of metabolism emphasizes mechanisms which require
interactions between enzymes in the control of metabolism. It is suggested that the PK-
CK interaction is such a mechanism, that it is tissue specific, and that it is specifically

adapted to the particular needs of skeletal muscle.

110

REFERENCES
Atkinson D. E. Cellular Energy Metabolism and its Regulation. (New York:
Academic. 1977.
Bailey J. E. Science. (1991) 252: 1668-1679.

Baranowska B., T. Baranowski. Molecular and Cellular Biochemistry. (1977) 17:
75-83.

Baranowska B., T. Baranowski. Molecular and Cellular Biochemistry. (1982) 45:
117-125.

Bessman S. P., R]. Geiger. Science. (1981) 211:448-452.

Blaedel W. J., T. R. Kissel, R.C. Bobuslaski. Analytical Chemistry. (1972) 44: 2030-
2037.

Brass J. M., C. F. Hiffins, M. Foley, P. A. Rugman, J. Birminghan, P. B. Garland.
Journal Bacteriology. (1986) 165: 787-94.

Brooks S. P. J ., K. B. Storey. American Journal Prysiology. (1988) 255: R289-R294.
Brooks S. P. J ., K. B. Storey. Journal of Theoretical Biology. (1991) 149: 361-375.

Cardenas J. M., R. D. Dyson. The Journal of Biological Chemistry. (1973) 248(20):
6038-6044.

Choate G. L., L. Lan, T. E. Mansour. Journal of Biological Chemistry. (1985) 260:
4815-4822.

Clark J. F. Kinetics of Porcine Carotid Artery Brain Isoform Creatine Kinase in situ
and in vitro. (1990) Dissertation for the Degree of PhD. MSU.

 

Clarke F. M., P. Stephan, G. Huxham, D. Hamilton, D. J. Morton. European Journal
of Biochemistry. (1984) 138: 643-649.

Colombo, M. F., D. C. Rau, V. A. Parsegian. Protein solvation in allosteric regulation:
a water effect on hemoglobin. Science. (1992) 256: 655-659.

Cook R., I. D. Kuntz. Annual Review of Biophysics and Bioengeneering. (1974) 3:
95-126.

111

Comish-Bowden A. European Journal of Biochemistry. (1991) 195: 103-108.

Creighton T. E. Proteins. Structures and Molecular Properties. (New York: W. H.
Freeman and Company. c1984).

Demoss J. A. Biochimica et Biophysica Acta. (1962) 62: 279-293.
Dick D. A. T. Journal of Theoretical Biology. (1963) 4: 98-112.
Dillon P. F., J. F. Clark. Journal of theoretical Biology. (1990) 143: 275-284.

Dillon P. F., M. K. Weberling, S. M. LeTarte, J. F. Clark, P. R. Sears, R. S. Root-
Bemstein. Journal of Biological Physics. (1995) 21: 11-23.

Eigenbrodt E., W. Schoner. Hoppe-Seyler's Zeitschrift fur Physiologische Chemie.
(1977) 358: 1033-1046.

Entman M. L., K. Kenichi, M. Goldstein, T. E. Nelson, E. P. Burnett, T. W. Futch, A.
Schwartz. Journal of Biological Chemistry. (1976) 251: 3140-3146.

Fisher M. J. and P. F. Dillon. NMR in Biomedicine. (1988) 1: 121-126.
Gaertner F. H. Trends in the Biochemical Sciences. (March 1978) 3: 277.
Hackenbrock C. R. Trends in the Biochemical Sciences. (1981) 6: 151-154.
Hasset A., W. Blattler, J. R. Knowles. Biochemistry. (1982) 21: 6335-6340.

Hardin C. D., L. Raeymaekers, R. J. Paul. Journal of General Physiology 99. ( 1992)
21-40.

Hsieh P. S., R. S. Balaban. Magnetic Resonance in Medicine. (1988) 7: 56-64.

Hol W. G. J., P. T. van Duijnen, H. J. C. Berendsen. Nature. (June 1978) 273: 443-
446.

Hyde C. C., S. A. Ahmed, E. A. Padlan, E. W. Miles, D. R. Davies. Journal of
Biological Chemistry. (1988) 249(43): 17857-17871.

Kay B. K., J. 1. Paul. Molecular Diversity. (1996) 1(2): 139-140.

Katz J., R. Rognstad. Current Topics in Cell Regulation. (1976) 10: 238-287.

112

Kohen E., G. R. Welch, C. Kohen, J. G. Hirschberg, J. Bereiter-Hahn. in The
Organization of Cell Metabolism. (G. R. Welch, J. S. Cleff. eds.) (New York:
Plenum. 1987) 251-275.

Kurganov B. 1., N. P. Sugrobova, L. S. Mil'man. Journal of Theoretical Biology.
(1985) 116: 509-526.

Liao J. C., Y. P. Chao, R. Patnaik. Annals of the N Y Academy Of Sciences. (1994)
745: 21-34.

Luther M. A., J. C. Lee. Journal ofBiological Chemistry. (1986) 261: 1753-1759.

MacDonald M., C. Chang. Molecular and Cellular Biochemistry. (1985) 68: 115-
120.

Marmillot P.,J-F. Hervagault, G. R. Welch. Proceedings of the National Academy of
Science. USA. (1992) 89: 12103-12107.

Matchett W. M. Journal of Biological Chemistry. (1974) 249: 4041-4049.
Melendez-Hevia E., F. Montero. Journal of Theoretical Biology. (1991) 152: 77-79.

Micheva K. D., A. R. Rarnjaun, B. K. Kay, P. S. McPherson. FEBS Letters. (1997)
414(2): 308-812.

Newsholme E. A., C. Start. Regulation in Metabolism. (1981).

Newsholme E. A., R. A. J. Chaliss, B. Crabtree. Trends in the Biochemical Sciences.
(1984) 9: 277-280.

Northrup, S. H. North Dakota EPSCoR Conference on Protein-Protein Interactions;
Physical and Kinetic Analysis of Protein-Protein Interactions. (May 1998)

Oberfelder R. W., B. G. Barisas, J. C. Lee. Biochemistry. (1984) 23: 3822.

Ottaway J. H., J. Mowbray. Physiological Reviews. (1976) 8: 107-208.

Ovadi J. Special issue: Physiological significance of metabolite channelling, a review
by Judit Ovadi, with critical commentaries by many authors. Journal of
Theoretical Biology. (1991) 152(1).

Pagliaro L. News in the Physiological Sciences. (1993) 8: 219-223.

Paul R. J., M. Bauwer, W. Rease. Science. (1979) 206(4425): 1414-1416.

113

Paul R. J. American Journal of Physiology. (1983) Cell Physiology 13: C399-C409.

Perryman M. B., A.W. Strauss, T. L. Buettner, R. Roberts. Biochimica et Biophysica
Acta. (1983) 747: 284-290.

Racker E. Journal Cellular Physiology. (1976) 89: 697-700.

Sackmann E. in Biophysics. W. Hoppe, W. Lohmann, J. Mark], H. Ziegler, eds.
(Berlin: Springer-Verlag. 1983) 425-457.

SacktorB., E. Wormser-Shavit. Journal of Biological Chemistry. (1966) 241: 624-31.

Schlichting I., X. J. Yang, E. W. Miles, A. Y. Kim, K. S. Anderson. Journal of
Biological Chemistry. (1994) 269(43): 26591-26593.

Selivanov V. A., D. R. Zakrzhevskaya,B. N. Goldstein FEBS--Letters. (1994) 345(2-
3): 151-3.

Solomon A. K. Reflections on the membrane-mediated linkage between cation
transport and glycolysis in human red blood cells. In: Membrane Transport
Processes. J. F. Hoffman, ed. (New York: Raven, 1978) 1: 31-59.

Smolen P., J. Keizer. Biophysical Chemistry. (1990) 38: 241-263.

Srere P. A. Annual Review of Biochemistry. (1987) 56: 89-124.

Srere P. A., B. Mattiasson, K. Mosbach. Proceedings of the National Academy of
Science. US. (1973) 70: 2534-2538.

Srivastiva D. K., S. A. Bernhard. Current Topics in Cellular Regulation. (1988) 28:
1-67.

Stephan P., F. Clarke, D. Morton. Biochimica et Biophysica Acta. (1986) 873: 127-
135.

Stephanopoulos G., J. J. Vallino Science. (1991) 252: 1675-1681.

Stossel T. Pour la Science--éditionfranpaise de Scientific American. (1994) 205: 62-
70.

Termonia Y., J Ross. Proceeding of the National Academy of Science. USA. (1981)
78(6): 3563-3566.

Thomas D., G. Broun, E. Selegny. Biochimie. (1972) 54: 229-244.

114

Ureta T. The role of isozymes in metabolite channelling. Journal of Theoretical
Biology. (1991) 152: 81-84.

Van der Veen K. J., A. F. Willebrands. Clinica Chimica Acta. (1966) 13: 312-316.
vanLeeuwen J. W. Biochimica et Biophysica Acta. (1983) 743: 408-421.

Wallimann, T., T.C. Doetschman, H.M. Eppenberger. Journal of Cell Biology. (1983)
96: 1772-1779.

Wallimann T., T. Schnyder, J. Schlegel, M. Wyss, G. Wegmann, A. Rossi, W.
Hemmer, H. Eppenberger, A. Quest. Progress in Clinical and Biological
Research. vol. 315 Muscle Energetics. (R. J. Paul, G. Elzinga, K. Yamada, eds.)
(New York: Alan R. Liss, Inc.; 1989) 159-176.

Welch G. R., J. S. Eastergy Trends in the Biochemical Sciences. (1994) 19(5): 193-
197.

Wilson J. E. in Current Topics in Cellular Regulation. B. L. Horecker, E. R.
Stadtman, eds. (1980) 16: 2-54.

115

l‘lICHI RN

6 STATE UNIV. LIBRARIES
l | 1111111111111111111111111111111111 11
31293020486498