Identification of important amino acid residues in the active site of yeast F1-ATPase and in the beef heart F1-ATPase in...


Material Information

Identification of important amino acid residues in the active site of yeast F1-ATPase and in the beef heart F1-ATPase inhibitor protein
Physical Description:
xiv, 176 leaves : ill. ; 29 cm.
Schnizer, Richard Arthur, 1961-
Publication Date:


Subjects / Keywords:
Research   ( mesh )
Proton-Translocating ATPases -- genetics   ( mesh )
Proton-Translocating ATPases -- chemistry   ( mesh )
Proton-Translocating ATPases -- physiology   ( mesh )
Amino Acids -- genetics   ( mesh )
Amino Acids -- chemistry   ( mesh )
Amino Acids -- physiology   ( mesh )
Amino Acid Sequence   ( mesh )
Binding Sites   ( mesh )
Recombinant Proteins   ( mesh )
Mutagenesis, Site-Directed   ( mesh )
Phenotype   ( mesh )
DNA, Mitochondrial   ( mesh )
Saccharomyces cerevisiae   ( mesh )
Cattle   ( mesh )
Heart   ( mesh )
Escherichia coli   ( mesh )
Department of Biochemistry and Molecular Biology thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Biochemistry and Molecular Biology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1993.
Bibliography: leaves 151-175.
Statement of Responsibility:
by Richard Arthur Schnizer.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002292989
oclc - 49673906
notis - ALP6179
System ID:

This item is only available as the following downloads:

Full Text






I would like to thank the members of my supervisory
committee, Dr. David Silverman, Dr. Al Lewin, Dr. Tom O'Brien, and
Dr. Brian Cain. Their guidance and suggestions were essential for
planning the course of my research, particularly with respect to the
pH studies of the mutant ATPases and the interpretation of the FII
data. Dr. Lewin's gifts of the anti-a, anti-P and anti-Fi sera were
greatly appreciated. In addition, I have benefitted from
collaboration with Gino Van Heeke on the E. coli e and beef heart FII
projects; a sincere thank you goes to Gino and his technicians Regina
Shaw and Judy Couton. I must also thank my advisor and friend Dr.
Sheldon Schuster for providing me with the opportunity to pursue
research in the unique atmosphere of his laboratory. His eagerness
to discuss new ideas and his flexibility in solving problems combine
very well with his willingness to allow independent research. His
perpetually positive attitude in the face of morbidly disappointing
data is an example that I hope to be able to emulate. I will need to.
The number of friendly faces that have moved through the
Schuster laboratory in the last several years is far more than you
could ever fit on a dart board. I am particularly grateful for the
friendships I have had with Sue Hinchman, Shijie Sheng, Charlie
Troxel, and Gino Van Heeke. Ellen Walworth has been a good friend,
and I cannot say enough about the job she does running the

laboratory, she made my life a lot easier. Holly Gray has made the
last few months very interesting and I will see her in Key West on a
catamaran. I am grateful to many other students outside our
laboratory with whom I have been friends, especially Richard Coffey
who has been an excellent friend, roommate, and source of
biophysical information.
Finally, I have to thank my parents who have been patient
throughout this process. The further I go through life, the more
grateful I am for the stable family they provided and the foundation
they gave me.



ACKNOWLEDGEMENTS ....................................................................................... ii

LIST OF TABLES ................................................................................................ v i

LIST OF FIGURES ................................................. ...... ............................. .... vii

ABBREVIATIONS............................................................................................ x

ABSTRACT .............................................. ............................................... xiii


1 INTRODUCTION .................................................................................... 1
Structure ........................................... ................................................ 3
Fo Subunits ................................................................. ....... 3
F1 Subunits ......................................... ...................................... 6
A ssem bly ............................................ ........................................... 2 2
Catalytic Mechanism ......................................................................... 25
R regulation ........................... ........... ... ......... .............................. 2 8

MUTANT ATPASES...............................................................................3 2
Introduction.......................................... ......... ........ ............................. 3 2
M materials .............................................................................................. 4 2
M methods ...................................... ...... .................................................... 4 5
R esults.......................................... ..................................................... 5 3
Discussion................................. ........... ........................................... 7 3

Introduction................................................................... 84
M methods ............................................... ......................................... 8 5

Results ................................................................................................... 86
FI-ATPASES ............................................................................................... 108
Introduction ..................................... ............................................108
Materials and Methods.............................. ............................... 09
Results ..................................................... ....................................... 118
Discussion............................................................................................1 34

5 SUMMARY AND CONCLUSIONS......................................................... 4 1

REFERENCES................................................. ......................................... 151

BIOGRAPHICAL SKETCH.................................... .......................................... 176



1.1 Precursors and Mature Subunits of the Fi-ATPase
C om plex. ................................................................. .................................. 7

1.2 Mutations in the P subunit of E. coli Fi-ATPase............................ 1 9

2.1 Specific activities of chloroform extracted ATPases
from mutant yeast strains.....................................................................60

3.1 Km values for ATP Hydrolysis by Mutant ATPases....................87

3.2 Values for pKa of ionizable groups involved in the
catalytic mechanism of wildtype and mutant
A TPases. ........................................... .................................................. 9 6

3.3 Hydrolysis of various nucleotide substrates by
submitochondrial particles from mutant and wildtype
yeasts................................. ....... ................. ........................................ 9 8

3.4 Kinetic Constants for ATP and dATP Hydrolysis by Pure
Y east Fl. ................................................... ................................................... 10 0

4.1 Turnover dependence of recombinant FlI...................................123

4.2 The effects of alanine substitutions in the region
from R35 to L45 of the beef heart Fi-ATPase inhibitor
protein................................................ ................................................. 3 0


2.1 Consensus sequence forming part of a nucleotide
binding site ......................................................... .................................. 3 3

2.2 The hydrophobic cleft of adenylate kinase.....................................3 6

2.3 A model of the E. coli Fl-ATPase P subunit nucleotide
binding site. ............................................................................................... 3 7

2.4 Construction of pMF ........................................................................... 44

2.5 Plasmid pOK........................................................................................... 46

2.6 The gapped heteroduplex technique of site-directed
mutagenesis...................................... .............................................. 4 8

2.7 Aerobic growth of wildtype and mutant yeasts...........................54

2.8 The position 211 mutations are recessive to wildtype..............5 6

2.9 Mutation L203F is an intragenic suppressor of the
position 211 mutations .......................................................................... 57

2.10 Quantitation of a and 0 subunits in submitochondrial
particles from mutant and wildtype yeasts and soluble
yeast Fi-ATPase.............................................................................. 63

2.11 Quantitation of the a subunit in submitochondrial
particles from mutant and wildtype yeast.............................6.....6 4

2.12 Relative amounts of a subunit in yeast strains
AVY4-1 and SEY103 .................................... ................................. 65


2.13 The sensitivity of wildtype and mutant Fl-ATPases
to oligom ycin.............................. .... .. ........ ........ .............................. 6 7

2.14 The effect of 2-propanol on ATP hydrolysis by
mutant and wildtype ATPases ....................................................... 69

2.15 The effect of ethanol on ATP hydrolysis by mutant
and wildtype ATPases......................................................... ..................70

2.16 The effect of methanol on ATP hydrolysis by mutant
and wildtype ATPases.............................. ............................... 7 1

2.17 The effect of dimethylsulfoxide on ATP hydrolysis
by mutant and wildtype ATPases........................................................72

3.1 Bicarbonate stimulation of FoFi-ATPase from strain
PH 21 N ...................................................................................................... 89

3.2 The influence of pH on ATPase activities of
submitochondrial particles from wildtype and strain
H 211K ....................................................................................................... 9 1

3.3 The influence of pH on ATP hydrolysis by
submitochondrial particles from mutant and wildtype
yeasts.......................................................................................................... 9 2

3.4 Plots of Vmax/Km for ATP hydrolysis by mutant and
wildtype ATPases as a function of pH...................................... ... 95

3.5 The influence of pH on the difference in transition
state binding energy between mutant ATPases and
the w ildtype. ........................................................................................... 1 05

4.1 Homology between Fi-ATPase inhibitor proteins
and the E. coli e subunit. .............................. .................................... 11 0

4.2 E. coli expression plasmid pALFII-6.................... ................ 11


4.3 Inhibition of E. coli FI-ATPase by purified recombinant
e and hCA H :e ....................................... .......................................... 20

4.4 Dixon plot of inhibition of E. coli F1-ATPase by e and
hC A ................................................................................................................12 1

4.5 Inhibition of beef heart and E. coli F1-ATPases by
recombinant FII .................................................................................... 124

4.6 Dixon plot of inhibition of beef heart Fi-ATPase by FiL...........125

4.7 The effect of ionic strength on inhibition of beef heart
ATPase activity by F ................................................ ............. 126

4.8 The effect of pH on inhibition of E. coli Fi-ATPase
activity by recombinant e subunit .................................................1. 28

4.9 Lineweaver-Burke plots of ATP hydrolysis by beef
heart F1-ATPase in the presence of varying amounts of
F I. ............................... ........ ........................................ 13 1

4.10 Lineweaver-Burke plot of MgITP hydrolysis by beef
heart Fi-A TPase.......................................................................................... 132

4.11 Lineweaver-Burke plots of ITP hydrolysis by beef
heart Fi-ATPase with varying amounts of FII..........................133

4.12 Dixon plot of inhibition by F1 inhibitor protein
mutant E40A. .........................................................................................1 35

4.13 Helical wheel arrangement of residues 32-49 of beef
heart FI-ATPase inhibitor protein....................................................138

ADP, adenosine-5'-diphosphate
amino acids are abbreviated according to the one letter code and the
three letter code
ATP, adenosine-5'-triphosphate
BCA, bicinechoninic acid
BES, N,N-bis(2-Hydroxyethyl)-2-aminoethanesulfonic acid
BICINE, N,N-bis(2-Hydroxyethyl)glycine
bp, base pair
C-terminus, carboxyl terminus
CD, circular dichroism
CHES, 2-(N-Cyclohexylamino)ethanesulfonic acid
CTP, cytidine-5'-triphosphate
dATP, 2'-deoxyadenosine-5'-triphosphate
dGTP, 2'-deoxyguanosine-5'-triphosphate
dCTP, 2'-deoxycytodine-5'-triphosphate
dTTP, 2'-deoxythymidine-5'-triphosphate
DEAE, diethylaminoethane
DMSO, dimethylsulfoxide
EDTA, ethylenediaminetetraacetic acid
EF1, Escherichia coli Fi-ATPase
EGTA, ethylene glycol-bis-(P-aminoethyl ether) N,N,N',N'-tetraacetic

FII, Fi-ATPase inhibitor protein
FTIR, Fourier transform infrared spectroscopy

g, gravity
GDP, guanosine-5'-diphosphate
GTP, guanosine-5'-triphosphate
hCA, human carbonic anhydrase isozyme II
hCA:e, human carbonic anhydrase isozyme II E. coli e subunit fusion

IDP, inosine-5'-diphosphate
IPTG, isopropylp-D-thiogalactoside

ITP, inosine-5'-triphosphate
kb, kilobase pair

Kcat, catalytic rate constant
kd, dissociation constant
Ki, inhibition constant
Km, Michaelis constant
M, molar
MES, 2-(N-morpholino)ethanesulfonic acid

Mr, relative molecular mass
N, normal
N-terminus, amino terminus
NADH, nicotinamide adenine dinucleotide
NMR, nuclear magnetic resonance
OSCP, oligomycin sensitivity conferring protein
PABA, para-aminobenzamidine

PAGE, polyacrylamide gel electrophoresis
PEP, phosphoenolpyruvate
Pi, inorganic phosphate
PIPES, piperazine-N,N'-bis(2-ethanesulfonic acid)

pKa, negative log of the acid dissociation constant
PMSF, phenylmethanesulfonylfluoride
PVDF, polyvinylidenedifluoride
SDS, sodium dodecylsulfate
TNP-ATP, 2'(3')-O-(2,4,6-trinitrophenyl) adenosine 5'-triphosphate
TRICINE, N-tris(hydroxymethyl)methylglycine
Tris, tris(hydroxymethyl)amino methane
UTP, uridine-5'-triphosphate

Vmax, maximum velocity
XTP, xanthosine-5'-triphosphate

Abstract of Dissertation Presented to the Graduate School of the
University of Florida in Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy

May, 1993
Chairman: Dr. Sheldon M. Schuster
Major Department: Biochemistry and Molecular Biology

Directed amino acid substitutions were constructed for H211 of
the Saccharomyces cerevisiae Fi-ATPase P subunit gene (ATP2). The
mutations (H211N, H211D, H211I, H211K, and H211A) were
expressed separately in atp2::LEU2 hosts and the resulting FOF1-
ATPases characterized. Yeast expressing mutation H211N showed
respired aerobically; the other mutant strains did not. These
phenotypes were suppressed intragenically by the mutation L203F,
implying a physical or functional interaction between residues 203
and 211. This supports a three-dimensional model placing L203 and
H211 in proximity. The mutant ATPases were unstable and could
not be purified. Compared to wildtype, mutant ATPases were
inhibited by ethanol and 2-propanol, showed reduced oligomycin
sensitivity, had altered nucleotide specificities, elevated Kms for ATP

hydrolysis, and low pH optima. Enzymes from mutant strains
H211N, H211D, and H211K showed substantial decreases in the pKa
of the catalytic general base relative to wildtype. This work
demonstrates that H211 and F203 are important for the stability of
the wildtype enzyme complex. H211 is required for the proper
assembly of FIFo, and also contributes to the structure of the active
site although it is not required in the catalytic mechanism and its
role is probably not dependent upon its ionization state.
In collaborative studies, recombinant E. coli F1-ATPase e

subunit and recombinant beef heart FI-ATPase inhibitor protein

(FII) were evaluated. Recombinant FII inhibited beef heart F1-
ATPase noncompetitively (Ki = 0.2 gM) in a turnover- and pH-

dependent fashion, and had no effect on E. coli F1-ATPase. The
amphiphilic a-helical region from R35-L45 of Fil was targeted by

alanine scanning mutagenesis. Substitutions in the proposed
hydrophobic face of this helix interfered with inhibition by FII while
mutations on the hydrophilic face had little effect. It was concluded
that specific hydrophobic interactions are required for inhibition by
FII. Recombinant e inhibited E. coli Fi-ATPase noncompetitively (Ki =

11 nM) in a turnover- and pH-independent fashion which was also
independent of the presence or absence of an N-terminal 30 kd
fusion protein. It was concluded that the N-terminus of e does not
interact with FI and that Fil and E inhibit their respective ATPases

through different mechanisms.

Plants and other photosynthetic organisms transduce
electromagnetic energy into the chemical energy which is used in the
biochemical processes of virtually all living things. This energy is
transferred between organisms primarily as carbohydrates, fats,
amino acids or proteins. Within cells chemical energy is stored in a
variety of molecules. The processes of metabolism require the
transfer of energy between these molecules. More of these transfers
are mediated by the high energy compound adenosine-5'-
triphosphate (ATP) than by any other molecule. ATP is one of many
cellular compounds containing high energy phosphate bonds which
serve as a source of free energy for the catalysis of reactions which
would otherwise be thermodynamically unfavorable. The hydrolysis
of ATP has an intermediate AG value relative to the other high

energy phosphate compounds common in biological systems. This
allows ATP and its hydrolysis products to donate and receive
phosphate bond energy and thereby to act as an energy shuttle.
The FoFI ATPase/synthase is a multisubunit transmembrane
protein which couples the energy of an electrochemical gradient to
the synthesis of ATP in the final step of oxidative phosphorylation.
The FO portion of the complex spans the inner mitochondrial
membrane and functions to conduct protons across it. This process

results in the collapse of the transmembrane electrochemical
potential created by the electron transport chain. The subunits of F1
and FO act together to couple the energy released by the transport
process to the synthesis of ATP by Fl. The precise mechanism of
energy transduction has not yet been determined and its elucidation
is the ultimate goal in the study of oxidative phosphorylation. The
enzyme is ubiquitous and is found in mitochondria, chloroplasts and
bacteria. It has been estimated that the most frequently occurring
enzyme catalyzed reaction on earth is the synthesis of ATP by FoFi
ATPases (Senior, 1988). Accordingly, the study of this enzyme
constitutes a major field of biochemistry, and the volume of
literature concerning this topic is substantial. A brief general review
is provided by Boyer (1987), and more detailed general reviews by
Senior (1988) and Kagawa (1984). Recent reviews focusing on
relationship of structure and function and on the role of symmetry
are also available (Cross, 1988, Ysern, Amzel, and Pedersen, 1988,
and Tiedge and Schafer, 1989). For reviews dealing more closely
with the enzyme mechanism see Futai et al., (1989), Boyer (1989),
Penefsky and Cross (1991), and Hatefi and Matsuno-Yagi (1992).
Extensive reviews dealing specifically with the E. coli enzyme have
been written by Senior (1990) and Fillingame (1990). This brief
introduction will summarize aspects of the structure, assembly, and
catalytic mechanism of the FoF1-ATPases with an emphasis on Fi.

Capaldi and coworkers have used electron microscopy to
describe the molecular architecture of the E. coli FoFI-ATPase (Gogol
et al., 1987, Gogol et al., 1989a, Gogol et al., 1990, Aggeler et al.,
1992a). The results of these authors are consistent with early reports
depicting the gross structure of the mitochondrial enzyme (Soper et
al., 1979, and Tzagoloff and Meagher, 1971). The F1 moiety appears
to be a 9 nm by 11 nm knob-like structure which is separated from
the membrane spanning FO portion by a narrow stem 4.5 nm in
length and 2 nm wide. FI can be stripped from Fo and exist as a
soluble complex capable of catalyzing the hydrolysis of ATP but not
its synthesis. Soluble F1 from rat liver mitochondria has been
crystallized and X-ray mapped to a resolution of 3.6 angstroms
(Bianchet et al., 1991). This complex is 7.6 nm "high" and 12 nm
wide when viewed on a plane which would be parallel to the
membrane, and 12 nm in diameter when viewed on a plane
perpendicular to the membrane. The subunit structure of FOFI is
discussed below.
E Subunits
In yeast (Devenish et al., 1992) and bovine (Hatefi and
Mitsuno-Yagi, 1992) mitochondria, FO is made up of six and seven
different subunits, respectively. E. coli FO contains only three
different subunits, a, b, and c (Foster and Fillingame 1979). Subunits
6, 8, and 9 of yeast FO and subunits 6, b, and c of bovine FO appear to
be equivalent to E. coli a, b, and c, respectively, and are thought to

form the transmembrane proton channel. The precise functions of
the remaining yeast and bovine subunits are unknown, although it
seems likely that some or all of them should be involved in forming
the stalk which joins Fo and F1. In yeast, these subunits have been
referred to as the FA complex by Nagley (1988) and are considered
to be separate from Fo but still involved in the coupling of proton
translocation to ATP synthesis.
The picture in E. coli is clearer and simpler. The
transmembrane pore through which protons are conducted consists
of one copy of subunit a having 5 to 8 transmembrane helices, and 9
to 12 copies of subunit c, each containing two transmembrane helices
joined by a polar loop (Foster and Fillingame, 1982, Senior and Wise
1983, Cain and Simoni 1989, Walker et al., 1984, Lewis et al., 1990).
There are also two copies of subunit b per complex. Each b subunit is
predicted to have a single transmembrane helix and two long
hydrophilic helices which project toward FI and may, together with
subunit 8, form part of the stem structure (Porter et al., 1985). All

three FO subunits are required for the assembly of a functional
proton pore (Schneider and Altendorf, 1985).
The a subunit (Mr 30,285) has been extensively mutagenized
and it appears that the C-terminal third of the protein is involved in
proton conduction. In particular, three residues, R210, E219 and
H245, (each protonatable) have been identified as crucial for this
process (Cain and Simoni 1986, 1988, 1989). Single mutants aR210,
aE219H and aH245E have 0%, 20% and 45% of wild type proton

translocation efficiency, respectively. Interestingly, the double
mutant aE219H,H245E has about 50% of wild type proton
translocation efficiency. This partial suppression of the translocation
defect was interpreted as evidence of a functional interaction
between residues 219 and 245, and also as evidence of physical
proximity. Deletions of hydrophilic domains in the N-terminus of the
a subunit show that the first 35 to 60 residues are important for
membrane targeting and insertion (Lewis and Simoni, 1992).
Subunit b has been overexpressed and purified (Dunn, 1992).
The relative molecular mass of this protein is 17,202. Chemical
crosslinking studies combined with sedimentation equilibrium, size
exclusion chromatography, and circular dichroism measurements
suggest that the purified b subunit forms an a helical dimer in
solution. The dimer is capable of inhibiting the binding of Fl to
naked Fo indicating an interaction with soluble Fl. Proteolysis of F1-
depleted membrane vesicles results in the removal of the
hydrophilic portion of subunit b. This does not affect FO proton
translocation (Perlin, et al., 1983), but does eliminate F1 binding in
reconstitution experiments (Steffens et al., 1987), lending further
support to the idea that the b subunit binds F1.
Subunit c is a small protein (Mr 8,264) which, together with
subunit a, is thought to compose the transmembrane proton pore.
Modification of c with dicyclohexylcarbodiimide (DCCD) occurs at
residue cD61 and results in a blockage of proton translocation (Hoppe
and Sebald, 1984). This charged residue is predicted to be located in

the middle of one of the two transmembrane helices of c. Mutations
of this residue to uncharged amino acids have also eliminated proton
transduction (Hoppe et al., 1980, 1982). These data suggest that
cD61 may be involved directly in the translocation of protons
through FO. It is interesting that DCCD modification of only one of the
9 to 12 c subunits present in Fo is required to inactivate proton
translocation (Hermolin and Fillingame, 1989). This is suggestive of a
cooperative interaction between c subunits. If proton translocation
involves the interaction of a single a subunit with each of the c
subunits, then a proton relay model can be envisioned where the c
subunits bind protons on the outer membrane side and transfer
them into the center of the membrane spanning region. At this point
the protons can be handed off to the a subunit and subsequently
transferred through the membrane. This type of model would
require the a subunit to be surrounded by c subunits and would also
require either rotation of the a subunit within the c complex or
rotation of the c complex around a.

FI Subunits
Fl-ATPases uniformly consist of nine subunits of five different
types. The subunits are named in order of decreasing Mr, and are
present in the complex in the ratio 3 a: 3 P: 7: 8: e (TABLE 1.1). The F1
inhibitor protein (FII) is loosely associated with F1 and will copurify
with it at neutral pH. This subunit will be discussed in a later section
in the context of catalytic regulation.

Table 1.1

Precursors and Mature Subunits of

the FI-ATPase Complex.

Subunit P M Mr P M M M Mr
a 580 544 58,500 553 510 55,164 509 55,200
3 511 492 54.575 528 480 51,595 459 50,155
y ? 309 34,000 299 272 30,141 286 31,428
8 ? 121 11,000 168 146 15,065 177 19,328
e 62 61 8,600 51 50 5,652 132 14,920
OSCP 212 195 20,870 213 190 20,968
FlI ? 63 7,383 109 84 9,572 _

The length of precursor and mature subunits is given as the total
number of amino acids. P: subunit precursor, M: mature subunit, Mr:
relative molecular mass, OSCP: oligomycin sensitivity conferring
protein, FII: mitochondrial F1 ATPase inhibitor protein. E. coli 8 is
equivalent to mitochondrial OSCP. E. coli E has sequences which are
homologous to both mitochondrial 8 and FII.

Yeast Mitochondria

Bovine Mitochondria

E. coli

The results of chemical crosslinking (Satre et al.,1976) and
electron micrographic studies (Tiedge et al., 1985) suggest that the a
and P subunits alternate with each other in a hexagonal arrangement.
This arrangement has been confirmed by the 3.6 angstrom crystal
structure of rat liver F1 (Bianchet et al., 1991) which shows two
slightly offset layers of three subunits each, one layer composed of a
subunits and one composed of 3 subunits. In each layer the subunits
are arrayed about a 3-fold axis of symmetry, the center of each
subunit separated from the others by 1200. The dimensions of the
individual subunits are 4.8 X 4.8 X 5.0 nm for a and 4.0 X 4.8 X 5.0
nm for p. The two layers interdigitate so that the "height" of the
complex is only 7.4 nm instead of 10 nm. Electron microscopic
studies show a central cavity in the hexagon which is occupied by an
off-center mass which is thought be composed of the y subunit and
either or both of the other minor subunits 8 and e (Tiedge et al.,

1983, Boekema et al., 1986, Gogol et al., 1989b). The crystal
structure of rat FI shows no central cavity. This may be an artifact
of the imaging process, e.g. the image of the asymmetrically located
mass is averaged out and consequently fills the central cavity.
Heavy metal labeling of cysteine residues supports this
interpretation. There are seven cysteine residues in rat FI, two in
each a subunit and one in y, but there are nine heavy metal atoms
detected in the X-ray map. Significantly, three of these atoms are
symmetrically arrayed near the center of the complex where
electron micrographs show the central mass to be. If the y subunit is

asymmetrically positioned in the central cavity, and an average
image of thousands of crystals is generated, then three heavy metal
atoms would appear symmetrically arranged in the center of the
hexagon. It should be noted that the authors considered this
explanation of their data to be unlikely. They expected the minor
subunits to appear as areas of low or discontinuous electron densities
if at all, and they fail to detect such areas. The issues of whether or
not the minor subunits compose the mass in the center of
mitochondrial FI, and which subunit owns the extra cysteines await
resolution by further research.
Sequence homology between F1 subunits from a variety of
different organisms is relatively high (Walker et al., 1985) reflecting
an evolutionary conservation of structure and function. This is not to
say there are no substantial differences between enzymes from
different organisms. For example, mitochondrial and E. coli enzymes
differ in their Fo sectors, as discussed above, as well as in FI. The a,
3 and y subunits of these enzymes are fairly homologous, and the E.
coli 8 subunit seems to correspond to mitochondrial oligomycin
sensitivity conferral protein (OSCP). However, the E. coli e subunit
has sequence homology with both the mitochondrial 5 subunit and
the loosely associated mitochondrial inhibitor protein (FiI), and there
is no known E. coli equivalent to the mitochondrial e subunit.
Sequence similarity also exists within F1 between the a and 1
subunits which appear to have evolved from a common ancestor
(Walker et al., 1982, 1985).

FI a subunits from various species generally range between 55
kd and 58.5 kd and are present in three copies per FI complex.
These subunits contain a consensus nucleotide binding site including
a GX4GKT/S motif, and have been shown to bind ATP very tightly
with no requirement for Mg2+(Cross and Nalin, 1982, and Issartel et
al.,1986). The characteristics of the catalytic nucleotide binding sites
include an affinity for a variety of nucleoside triphosphates, a
requirement for a divalent cation, and free exchange with medium
nucleotides. Because the a subunit binding site is specific for ATP,
does not exchange nucleotides freely with the medium, and has no
Mg2+ requirement, it cannot be a catalytic site. The actual role of the
a subunit nucleotide binding site is unknown at this time, although a
recent report suggests that binding of nucleotides to noncatalytic
sites is required for catalytic positive cooperativity and is
responsible for negative cooperativity of substrate binding (Jault and
Allison, 1993). It is not yet clear whether this noncatalytic site is
actually associated with the a subunit. It is also not certain that the
nucleotide binding site attributed to the a subunit is contained

entirely within that subunit. It has been suggested that it is actually
a shared site, located at the interface of a and P subunits and

composed of sequences from each (Kironde and Cross, 1987, and
Wise et al., 1987).
The relationship of a subunit structural features to function has
been reviewed by Senior (1990). The N-terminal portion of E. coli a
has been implicated in binding the 8 subunit (homologous to

mitochondrial OSCP) to F1 (Dunn et al., 1980 and Maggio et al.,
1988). When the N-terminal 15 amino acids are removed from a,
the FI complex does not bind 8 and consequently cannot bind to F0.
On the basis of sequence homology to other nucleotide binding
proteins (Walker 1982), mutational analyses (Rao et al., 1988b), and
labeling by nucleotide analogues (Wise et al., 1987), the nucleotide
binding site has been assigned to the region of residues 160-340.
Senior (1990) has proposed that the segment of the a subunit from
residues 345 to 375 is part of an interfacial surface with the P
subunit, and that this region is involved in transmitting
conformational changes between a and P subunits. Mutations in this
region have been shown to diminish both positive catalytic
cooperativity and negative cooperativity of substrate binding, while
having no effect on unisite catalysis (Maggio et al., 1987). Jault et al.,
(1991) report a mutation at position E173 in the nucleotide binding
region of the a subunit which also diminishes negative cooperativity.
This mutation is far outside the so-called signal transmission region
but may still be at an interface between a and P if the nucleotide
binding site is shared. In this case it is conceivable that this residue
could be involved in direct intersubunit communication. One other
potential function of the a subunit is its possible role as a chaperone
to the 1 subunit. This role is discussed below in the context of
The E. coli FI y subunit has been shown by electron microscopy
to occupy the center of the hexagonal array of a and 3 subunits

(Gogol et al., 1989b). The y subunit is required for assembly of Fl in
vivo (Miki et al., 1988), and in vitro (Dunn and Futai, 1980) and has
been crosslinked to both the a and 0 subunits (Aggeler and Capaldi,
1992) as well as to e (Aggeler et al., 1992b). Purified F1 subunits
from either E. coli or thermophilic bacterium PS3 can be reassembled
to form functional ATPases (Futai, 1977, and Yoshida et al., 1977). In
E. coli FI the y subunit is required for functional reconstitution, the
minimal complex required for activity being as333y. PS3 subunits will
form several functional oligomers including a3s3, a3133, and a3038
(Yokoyama et al., 1989). Of these, the as3y complex has the greatest
activity. The significance of these observations may simply be that
the y subunit is required for conformational stability of the a/P
hexamer, or perhaps it is involved in the communication of
conformational changes. The function of y is not yet clear although it
appears to be involved in energy coupling in the FoF1 complex. Shin
et al. (1992) have isolated mutations at position 23 of the E. coli y
subunit (yM23R and yM23K) which disable proton pumping without
affecting ATPase activity. Nakamoto et al. (1993) have generated
several suppressor mutations of yM23R and yM23K by random
mutagenesis of the entire unc G coding sequence. The suppressor
mutations alter amino acids located in the C-terminus of y primarily
in the region from Q269 to V280. The authors argue that
suppression implies a physical interaction between the regions
surrounding the mutated residues. This serves as a basis for a model
of a portion of the y subunit. Secondary structure predictions suggest

that M23 and the loci Q269-V280 are located in a-helices. The
authors claim that these helices are in proximity in the tertiary
structure of y, and that they both play a role in energy coupling.
The 8 subunit of mitochondrial ATPases bears sequence
similarity to the E. coli e subunit and the two are considered to be
equivalent (Walker et al., 1985). No function has been ascribed to
the mitochondrial 6 subunit yet, but crosslinking experiments show
that it is located near the 7 subunit in FoFI (Joshi and Burrows, 1990).
The E. coli E subunit is required for binding of F1 to FO (Sternweis,
1978) and may also regulate ATPase activity. E. coli e is not bound to
F1 as tightly as the other subunits and will dissociate from the
soluble enzyme with kd in the nanomolar range. Mitochondrial 8 is
bound more tightly and is not observed to dissociate from the
complex any more readily than the other subunits. Dunn et al.
(1987) have utilized monoclonal antibodies directed against E. coli
to completely deplete FI of the e subunit. A comparison of Kcat
values for ATPase activities of e-depleted and e-replete soluble F1
shows that the association of e with F1 decreases Kcat by about 6-fold
(Dunn et al., 1987). Soluble F1 is inhibited by epsilon in a
noncompetitive manner. Dunn et al. (1982), using e-depleted F1,
have calculated the Ki for e to be 0.3 nM. Previous studies using
enzyme only partially depleted of e reported ki in the 10 nM range
(Smith and Sternweis, 1977, and Sterweis and Smith, 1980), thus
suggesting the possibility that e has a regulatory role in the catalytic
mechanism. E. coli mutants which express no e subunit grow very

poorly and have lower concentrations of cellular ATP than wild type
strains (Porter et al., 1983, Klionsky et al., 1984). On the other hand,
Sternweis and Smith (1980) have shown that reassociation of F1 with
Fo relieves the inhibitory effect of e, suggesting that inhibition by e
may be artifactual. However, Mendel-Hartvig and Capaldi (1991)
report that proteolysis of e in FIFO results in an increase in the rate of
ATP hydrolysis. Klionsky et al. (1984) have suggested that the
function of e is to inhibit the ATPase activity of F1 moeities not yet
bound to Fo. While this may be one of the functions of e, it seems
unlikely that it would be its only function. Why would the cell
evolve a protein which is a stable part of the assembled FoFI
complex, but whose only function is preventing premature ATP
hydrolysis by unassembled FI?. It would make more sense if it had
a role in FOFI as well. Mendel-Hartvig and Capaldi (1991) report that
the conformation of the e subunit changes when the enzyme binds
nucleotides in the catalytic binding site. They combine this evidence
with the fact that e decreases the rate at which Pi dissociates from
the enzyme (Dunn et al., 1987, and Wood et al., 1987), and propose
that the primary function of e is in coupling events at the catalytic
sites with proton pumping.
The mitochondrial e subunit is the smallest subunit of F1 with
Mr of 5652 in bovine heart and 8600 in yeast. Genes encoding the
mitochondrial e subunit from both bovine heart (Vinas et al., 1990)
and yeast (Guelin et al., 1993) have been cloned and sequenced. The
mature protein has been isolated from several sources including pig

heart (Penin et al., 1990, Gagliardi et al., 1991), beef heart (Walker et
al., 1985), yeast (Arselin et al., 1991), and sweet potato (Kimura et
al., 1989). The available primary structure information shows that
these proteins are homologous. The e N-terminus has the
characteristics of both a mitochondrial targeting sequence and a
leader sequence, although only the N-terminal methionine is cleaved
to form the mature subunit (Vinas et al., 1990, Guelin et al, 1993).
Joshi and Burrows (1990) have used homobifunctional crosslinking
reagents to show that beef heart e is located close to y in the FoFI
complex. The e subunit from pig heart has been purified by

reversed-phase high pressure liquid chromatography as a tightly
bound heterodimer with the 6 subunit (Penin et al., 1990, Gagliardi et
al., 1991). The purified e and 8 subunits will reassociate to form a
dimer which is indistinguishable by criteria of chromatographic
behavior, circular dichroism spectrum, and intrinsic fluorescence
from the e8 dimer purified directly from Fl. Intrinsic fluorescence
measurements indicate that the only tryptophan in e, which is
located in the N terminus, is involved in the interaction with 8.

Guelin et al. (1993) have constructed a yeast strain in which
the chromosomal copy of the e gene (ATPe) has been disrupted. The

resulting strain is constitutively anaerobic and has no oligomycin-
sensitive ATPase activity. The mutation effectively uncouples proton
transport from ATP synthesis. The absence of E causes a proton leak
in Fo, possibly by influencing FO conformation. The resulting F1
moiety is unstable and cannot be purified by extraction with

chloroform. The mutation is complemented by transformation of the
mutant strain with an expression vector for ATPe. These results
show that the e subunit of mitochondrial ATPase is required for
proper assembly and activity of the complex, in particular functional
coupling of F1 and FO requires the presence of e.
The data presented in the previous three paragraphs show that
the mitochondrial y, 8 and e subunits are all closely associated by
criteria of crosslinking or copurification. The E. coli y subunit is
homologous to the mitochondrial y subunit and has been located in
the center of the a/ct hexagon. It therefore seems almost certain that
the central mass detected in electron micrographs of F1 is composed
of one or more of the minor subunits.
The oligomycin sensitivity conferral protein (OSCP) is a
member of the mitochondrial FoFI-ATPase complex and is located in
the stalk region which connects FO to Fl (McClellan and Tzagoloff,
1968). OSCP facilitates the binding of FI to FO in mitochondrial
ATPases (Tzagoloff, 1970, and Penin et al., 1986) and is required for
binding in the case of yeast (Uh et al., 1990). It has been shown to
bind to Fo (Dupuis and Vignais, 1987) and to FI independently, and
has been crosslinked to the a and P subunits with zero-length
crosslinkers (Dupuis et al., 1985). The E. coli 8 subunit is homologous
to OSCP and appears to have a similar function and location within
the complex (Sternweis and Smith, 1977). FoFi complexes which
have been depleted of OSCP are incapable of ATP synthesis or Pi-ATP
exchange, but this situation can be reversed by the addition of

exogenous OSCP. Genes encoding bovine (Joshi et al., 1992) and yeast
OSCP (Uh et al., 1990) have recently been cloned. In the yeast
system the clone was used to construct a chromosomal deletion
strain which was incapable of aerobic respiration and which had no
detectable oligomycin-sensitive ATPase activity. The bovine protein
was expressed in E. coli, purified, and shown to have biological
activity. A series of C-terminal deletion mutants was then generated
by inserting stop codons 10, 20, 30, or 40 codons upstream from the
authentic stop codon. Each of the resulting proteins was expressed
and purified, and found to be ineffective in restoring ATPase or Pi-
ATP exchange. Unfortunately the authors did not convincingly show
that the mutant proteins were folded properly, so at this time there
is no information concerning the function of specific regions of the
F1-ATPase J subunits are present in three copies per complex
and contain one nucleotide binding site each. The characteristics of
these sites include a preference for adenine nucleotides, but an
affinity also for GTP, dATP, ITP, XTP, CTP, and UTP (Pullman et al.,
1960, and Schnizer and Schuster unpublished), a requirement for a
divalent cation (Pullman et al., 1960), and frequent exchange of
bound nucleotides with medium nucleotides (Cross and Nalin,1982,
and Issartel et al.,1986). These are properties which are
characteristic of catalysis by F1, and it is generally agreed that the P
subunits contain the catalytic nucleotide binding sites. This idea is
supported by the unusually high degree of sequence homology seen

between P subunits of various ATPases (Walker et al., 1982, 1985),
and by the fact that the isolated P subunits of Rhodospirillum
rubrum (Harris et al., 1985) and possibly E. coli (Al Shawi et al.,1990)
can catalyze ATP hydrolysis whereas isolated a subunits cannot. The
P subunit contains a GX4GKT/S motif and other regions of homology
to known nucleotide binding proteins (Walker et al., 1982).
Extensive mutagenesis of the P subunits of E.coli FI has helped to
identify specific amino acid residues and general regions of the
protein which are important for catalysis and assembly (reviewed in
Senior 1988, 1990, and Futai et al., 1989). Table 1.2 lists several E.
coli P subunit mutations and their consequences.
The use of nucleotide analogues capable of labeling nearby
groups has aided in identifying amino acids proximal to the
nucleotide binding site (Cross et al., 1987, Garin et al., 1986,
Hollemans et al., 1983, Bullough, and Allison, 1986). In addition,
sequence homology with other nucleotide binding proteins of known
structure such as adenylate kinase, ras p21, and EF-Tu has provided
a guide for the study of nucleotide binding by F1 P subunits. Walker
et al., (1982) have defined two regions of homology, referred to as A
and B, which are present in many nucleotide binding proteins. The A
region encompasses the GX4GKT/S motif and the B region includes an
aspartate residue which may be involved in stabilizing the complex
of the divalent cation with the nucleotide. Thomas et al., (1992) have
deleted these regions from cDNAs encoding the C-terminal three
fourths of the rat p subunit, expressed the resulting proteins in

Table 1.2

Mutations in the P subunit of E. coli Fi-ATPase.


C137S catalysis Kironde et al. (1989)
Y catalysis _
G142S catalysis, positive Parsonage et al. (1987)
D catalysis Kironde et al. (1989)
G146S catalysis Kironde et al. (1989)
G149S suppresses S174F Miki et al. (1990)
I assembly Senior & Al Shawl (1992)
G1491, G1541 "
G150A, A151G, T156S catalysis Takeyama et al. (1990)
A151P twice wild type activity Takeyama et al. (1990)
A151P, V153S, T156G severely inhibited catalysis Takeyama et al. (1990)
V catalysis _
G152D catalysis Lee et al. (1991)
R "
G1541 assembly Senior & Al Shawl (1992)
K155A catalysis Omote et al. (1992)

SQ severely inhibited catalysis Senior & Al Shawi (1992)
K155-G-T156 severely inhibited catalysis Takeyama et al. (1990)
T156A severely inhibited catalysis Omote et al. (1992)
"C ____
"S activity 50% greater than
wild type
E161Q catalysis Lee et al. (1991)

S174F Mg2+ catalysis, positive Noumi et al. (1984),
cooperativity, no effect on Parsonage et al. (1987)
Ca2+ catalysis
E181Q severely inhibited catalysis Senior & Al Shawl (1992)
E185Q assembly Noumi et al. (1987)
"K K .
E1920 catalysis Parsonage et al. (1988)
G207D catalysis Kironde et al. (1989),
Senior & Al Shawi (1992)
M2091 catalysis, positive Parsonage et al. (1987)

Table 1.2 Continued


G214R assembly Parsonage et al. (1987)
D242N severely inhibited catalysis Senior & Al Shawl (1992)
R246C catalysis, positive Parsonage et al. (1987)
G251D catalysis Lee at al. (1991)
T285D catalysis Noumi at al. (1988)
A295T suppresses S174 Miki et al. (1990)
P ___
Y297F none Wise (1990)
D301V assembly Lee et al. (1991)
D302V assembly Lee et al. (1991)
Y331S catalysis Wise (1990)
"F 50% wild type activity, Parsonage et al. (1987),
tyrosine-OH not required Wise (1990),
for catalysis Weber et al. (1992)
A catalysis Weber et al. (1992), Wise
E Weber et al. (1992)

Y354F none Parsonage et al. (1987)
R398H aurovertin resistant, no Lee et al. (1989)
effect on catalysis
R398W ___ _____Lee at al. (1991)
R398C __
L400Q suppresses S174 Miki et al. (1990)
P403SG415D assembly Senior & Al Shawl (1992)

E. coli, purified them, and assayed their affinity for the fluorescent
nucleotide analogue 2'(3')-O-(2,4,6-trinitrophenyl) adenosine 5'-
triphosphate (TNP-ATP). The affinities for TNP-ATP were compared
to that of an overexpressed and purified wild type P subunit. They
report that only the absence of the A region affects nucleotide
binding. The A region deletion mutant showed a 30-fold decrease in
affinity relative to wild type, while the B region deletion mutant
showed no change in affinity. Both of these proteins and the wild
type subunit were assumed to have folded correctly on the basis of
circular dichroism spectra and resistance to E. coli proteases.
Duncan et al. (1986) have proposed a three-dimensional model
for the nucleotide binding site of E. coli P which is based on the
crystal structure of adenylate kinase. These authors suggest that the
nucleotide binding site is contained within the protein segment from
G142 to P332 and includes six P strands which alternate with five a
helices in the primary structure. This model was presented as an
hypothesis and will remain hypothetical until a high resolution
crystal structure of the P subunit is available. However, there are
data which support its validity. Table 1.2 lists many mutations in
this region which have a negative effect on catalysis. Included in
this domain is the GX4GKT/S motif which forms a loop between the
first 3 strand and first a helix. The lysine residue in this loop is
thought to interact with the y-phosphate of ATP in the cases of
adenylate kinase (Pai et al., 1977) and p21 (Pai et al., 1989). An
involvement in catalysis has not been demonstrated, but mutations

of this lysine (Reinstein et al.,1988, Tian et al., 1990) and of the
corresponding one in ATPase P subunits (Table 1.2, K155) have
severe negative effects on catalysis without apparent structural
perturbation. Furthermore, the crystal structure of adenylate kinase
(Pai et al., 1977) shows that the adenine-ribose moiety of ATP is in a
hydrophobic pocket composed of five amino acid residues, three of
which are clearly conserved in FI P subunits. Weber et al., (1992)

used fluorescence spectroscopy to show that the environment
surrounding the adenine moiety in the catalytic site of E. coli Fl was
hydrophobic. These data certainly do not confirm the model but
they are consistent with it. Even though it remains unproven, the P
subunit nucleotide binding site model is useful in the design of
experiments and in the speculative interpretation of data.

In the case of mitochondrial enzymes, each of the FI subunits is
encoded by a nuclear gene and is expressed initially as a precursor
(Maccecchini et al., 1979, Lewin et al., 1980, Runswick et al.,1990,
Vinas et al., 1990). These precursors must be delivered to the
mitochondria, transported across two membranes, proteolytically
processed, and assembled into a functional enzyme. The N-termini of
the known precursors have the characteristics of a membrane
targeting and insertion sequence, including predicted amphipathic
secondary structure with positively charged residues opposite
hydrophobic ones.

The import pathway of each nuclearly encoded FoFI subunit is
not known in detail, but the P subunit has served as a model system
for mitochondrial protein import and it is likely that certain aspects
of its import pathway are similar to those of the other subunits. P
subunit precursors are folded in the endoplasmic reticulum and
transported to a mitochondrion at a contact site between the inner
and outer membranes. Association with Hsp70SSA and
uncharacterized cytosolic factors (Pfanner and Neupert, 1990, Hartl
and Neupert, 1990) results in an ATP-dependent unfolding of the
precursors which then interact with signal sequence receptors (eg.,
MOM19, Sollner et al., 1989). Each precursor is then transferred to a
general insertion protein (eg., ISP42, Vestweber et al., 1989) and
transported through the mitochondrial inner membrane by an
unknown apparatus. This transport is dependent on the existence of
an electrical potential across the inner membrane, the matrix having
a negative charge relative to the cytosol (Pfanner and Neupert,
1986). Signal sequences are cleaved by a mitochondrial matrix
processing peptidase (Hawlitschek et al., 1988) accompanied by a
processing enhancing protein (Yang et al., 1988). Mature subunits
are thought to be removed from the translocation apparatus by
hsp70SSCl (Kang et al., 1990 Scherer et al., 1990) which is associated
with the matrix face of the inner membrane. The subunits are then
refolded in an ATP-dependent process involving hsp60MIF4
(Ostermann et al., 1989). Subsequent assembly of P subunits into

macromolecular complexes also involves hsp60MIF4 (Cheng et al.,
It has been proposed that the FI a subunit acts as a chaperonin
for the P subunit (Luis et al., 1990, Avni et al., 1991). Yuan and
Douglas (1992) suggest that the a subunit is not a molecular
chaperone but rather an "assembly partner" which influences protein
import without influencing folding. They show that the a subunit is
necessary for efficient import of precursor proteins, particularly the
I subunit. The other FI subunits are imported more efficiently in
strains where the B subunit is imported normally than in those from
which it is absent or imported poorly. The a subunit therefore
indirectly influences the import of the y, 5 and e subunits. These
authors propose a model in which the a subunit aids in the discharge
of P from hsp70sSC1 and that this release is coupled to the immediate
assembly of P subunits into the complex. This is consistent with the
finding of Burns and Lewin (1986) that the P subunit is assembled
soon after import.
The issue of how the subunits of FO and FI assemble into a
functional complex has not been settled. The assembly of E coli FIFO
has been proposed to occur by either sequential addition of subunits
(Cox et al., 1981, Cox et al., 1987), or by association of all of the FI
subunits into the ATPase complex, and subsequent binding to
previously assembled and membrane-integrated Fo (Sternweis and
Smith, 1977). Evidence consistent with the second model has been
reported in yeast where functional F1 moieties are assembled, and

associate with the inner mitochondrial membranes of strains which
do not synthesize FO subunits 6, 8, and 9 (Orian et al., 1984). Lewin
and Norman (1983) have shown that imported FI subunits are
assembled into new complexes and do not replace subunits in
existing complexes. In the yeast system accessory proteins have
been identified which are involved in the assembly of FO and F1, but
which are not directly associated with the complex. The ATP10 gene
product is required for the proper assembly of FO (Ackerman and
Tzagoloff, 1990), and the ATP11 and ATPI2 gene products are
required for the assembly of FI (Ackerman et al., 1992, and Bowman
et al., 1991). It is probable that other proteins involved in the
assembly of FIFo will be identified in the future, and the complete
understanding of the assembly process awaits their discovery.
Catalytic Mechanism
As stated above, the ultimate goal of the study of oxidative
phosphorylation is to determine the mechanism by which the energy
of a transmembrane electrochemical gradient is transduced into
phosphate bond energy in the form of ATP. The process of
transferring protons across a membrane is facilitated by FO, and the
energy released in that process is somehow coupled to the catalytic
site on an FI f subunit. Simplistically there are two ways for this to
happen. One involves a direct coupling in which protons are passed
through F0 and FI to a catalytic site where they abstract oxygen from
Pi, thereby facilitating a nucleophilic attack by the P-phosphate of
ADP (Mitchell, 1974). The other way involves indirect coupling by

transmission of conformational changes. In this conformational
coupling system the protons transferred through FO cause
conformational changes in the Fo subunits which are then
transmitted to F1 subunits and finally to the catalytic site(s).
Transfer of protons into FI is not required because the
conformational changes in the catalytic sites drive the reaction
In the early 1970s it was noted that in the absence of any
protonmotive force the ATPase enzyme "idled." That is, both the ATP
synthesis and hydrolysis reactions occurred without product release.
This was discovered by 180 exchange experiments in which FI loaded
with unlabeled ATP was incubated in 180 labeled water. The ATP
was labeled with 180 without being released from the catalytic site
(Boyer et al., 1973). It turned out that the equilibrium constant for
the interconversion of ADP + Pi and ATP on the enzyme was nearly
1.0. This implied that energy input was not actually required to
drive the reaction chemistry. Boyer then proposed that the energy-
requiring steps in ATP synthesis or hydrolysis were actually the
binding of substrates and the release of products, and that these
processes depended on conformational changes in the catalytic sites
(Boyer et al., 1973, Kayalar et al., 1977). He proposed an alternating
site model of catalysis which came to be known as the three site
binding change mechanism. Two site versions of the model have also
been proposed wherein the third nucleotide binding site is
regulatory and not catalytic (Berden et al., 1992, Bullough et al.,

1987). A recent report from Zhou and Boyer (1993) on the
mechanism of photophosphorylation by chloroplast ATP synthase is
also consistent with either two-site or three-site models.
The binding change mechanism is cyclical and the three-site
version can be summarized as follows. Three catalytic sites exist in
F1, one on each 0 subunit. At a given time one of these sites has
tightly bound ATP, the second has tightly bound ADP and Pi in
equilibrium with ATP, and the third is empty and in a conformation
which will not bind substrates. Energy from the electrochemical
gradient is then released and causes a conformational change in Fo
subunits which is transmitted to Fl. Each of the binding sites then
undergoes a conformational change. The first site releases ATP and
adopts the empty site configuration. The second site binds ATP
tightly, and the third site adopts a conformation allowing it to bind
ADP and Pi tightly. The cycle can then repeat. An important aspect
of this mechanism is that it accommodates negative cooperativity of
substrate binding and positive catalytic cooperativity, both of which
are documented characteristics of catalysis by FoF1 (Grubmeyer and
Penefsky, 1981, Cross et al., 1982).
Another important implication of the mechanism is the
functional asymmetry of nucleotide binding sites. At no time do any
two sites have the same characteristics, and each site must pass
sequentially through each of the three possible conformational states.
This functional asymmetry is paralleled by the structural asymmetry
seen in electron micrographs (discussed above) in which a central

mass is associated with one a/P pair. Cryoelectron microscopy
studies using monoclonal antibodies to E. coli e and a subunits have
shown that when ATP is bound in the absence of Mg2+ (a
noncatalytic condition), the e subunit associates with a P subunit
independently of the location of the central mass (Gogol et al., 1990).
However, in the presence of Mg2+ and ATP (under the conditions of
catalysis), the central mass and E subunit are usually associated with
the same P subunit. These observations are consistent with the idea
that the minor subunits are involved in the communication of
conformational changes involved in catalysis.
The alternating positions occupied by the a and 0 subunits in
the Fj complex suggest that the a subunits are also involved in
intersubunit communication, and therefore may play a role in the
sequential changes in the conformation of each binding site. This
idea is supported by mutational analyses of the a subunit. Senior's
laboratory has isolated a subunit mutants which inhibit multisite
catalysis (positive cooperativity) as well as negative cooperativity of
substrate binding without affecting the total number of nucleotides
bound (Wise et al., 1984). It seems clear that the catalytic
mechanism of F1-ATPase is extremely complex and, even though
catalysis occurs on the P subunits, the kinetic characteristics depend
on the interactions of all of the F1 subunits.

The mitochondrial ATP synthesis reaction is reversible in vivo.
In constitutive anaerobes the ATPase functions as a proton pump.

The enzyme uses ATP to generate a membrane potential which is
ultimately used for nutrient uptake. Mitochondria also run the
hydrolysis reaction in order to reduce oxidized electron carriers.
This information raises the issue of regulation. Biological systems
need a mechanism for regulating the activity of the FoFI-ATPase in
times where there is a need for a large cellular pool of ATP. If such a
pool is accumulated at the expense of the proton gradient, and the
concentration of ATP is far greater than that of ADP, then there
needs to be a mechanism of avoiding a futile cycle. In the case of
mitochondrial ATPases this regulation is achieved through the action
of an Fi-ATPase inhibitor protein (FII). Inhibitor proteins have been
discovered in a variety of sources including beef heart (Pullman and
Monroy, 1963), rat liver (Cintron and Pedersen, 1979)
Saccharomyces cerevisiae (Hashimoto et al., 1981), and Candida utilis
(Klein et al., 1977). Proteins which may have a similar function
include the e subunits of E. coli (Smith and Sternweis, 1977) and
chloroplasts (Nelson et al., 1972, and Nieuwenhuis and et al., 1974).
The interaction of FII with FI has been best characterized in
the beef heart system. Bovine F11 has been shown to bind to the P
subunit of FI in a stoichiometry of one per FI complex (Klein et al.,
1980, 1981, Wong et al., 1982), so it is likely that inhibition of ATP
hydrolysis is mediated through a single P subunit. Current models of
the interaction of the FII with Fi.ATPase depend on conformational
changes in the inhibitor protein itself (Panchenko and Vinogradov,
1985, Fujii et al., 1983, Milgrom, 1991), and possibly the ATPase

complex (Schwerzman and Pedersen, 1986). The inhibitor protein
can exist in at least two conformations, active and inactive. The
active conformation predominates at acidic pH, and the inactive one
at basic pH. Protonation of histidine residues has been implicated in
the conformational transition of both bovine and yeast inhibitor
proteins (Panchenko and Vinogradov, 1985, Fujii et al., 1983). The
active form of the inhibitor inactivates Fi immediately, whereas
inhibition by the inactive form takes several minutes and is
dependent on catalytic turnover (Schwerzman and Pedersen, 1986,
Panchenko and Vinogradov, 1985). Conversion between active and
inactive forms depends on pH in vitro, and is thought to depend on
the state of mitochondrial energization in vivo (Van de Stadt et al.,
1973, Harris et al., 1979, Husain and Harris, 1983). The inhibitor
assumes its active form in mitochondria under nonenergizing
conditions and reverts to its inactive form upon formation of the
proton gradient. Milgrom (1991) has shown that FI inhibited by Fil
contains a single nucleotide trapped in a catalytic site. According to
the binding change model, once one catalytic site is emptied by
release of products an alternate catalytic site assumes the
conformation required for product release (Kayalar et al., 1977).
Milgrom (1989, 1991) suggests that the inhibitor protein functions
by blocking product release at a single catalytic site, and that this
eliminates the cooperative interactions necessary for promoted
catalysis. This idea is supported by experiments showing that the

FII-F1 complex is capable of unisite noncooperativee) ATP hydrolysis
(Vasquez-Laslop and Dreyfus, 1990).
The picture that has been presented in this introduction is one
of an extremely complex and important multisubunit enzyme. The
work in the chapters that follow concentrates primarily on the role of
a single amino acid in the 0 subunit of the yeast FoFi-ATPase. The
data presented will show that this amino acid plays a role in
maintaining the correct conformation of the P subunit and that it is
essential for normal assembly and normal catalysis. Further data
will provide support for an adenylate kinase-like configuration of the
nucleotide binding site. The last chapter of this work focuses on the
regulation of the beef heart Fi-ATPase. A collaborative study was
undertaken to identify important amino acid residues in the beef
heart F1 inhibitor protein as well as to compare some of its
characteristics to those of the inhibitory e subunit of E. coli.



The FoFI ATPase/Synthases catalyze the formation of ATP in
the final step of oxidative phosphorylation. The F1 portion of the
enzyme contains the catalytic domains and consists of five different
subunits in an a3:13:y:8:e stoichiometry. The catalytic nucleotide
binding sites are found on the P subunits (Cross and Nalin, 1982).
The structural constituents of these sites have not yet been
rigorously identified due to the technical difficulty in obtaining high
resolution NMR or X-ray crystallographic information for such a
complex enzyme. However, analysis of the P subunit primary
structure reveals a consensus sequence found in many other
nucleotide binding proteins (Fry et al., 1986, Walker et al., 1985)
(Fig.2.1). This sequence has been shown to form part of the
nucleotide binding site in the enzymes adenylate kinase and ras p21.
High resolution tertiary structures have been solved for both of these
enzymes in the nucleotide-bound state (Fry et al., 1985, Fry et al.,
1988, De Vos et al., 1988, Pai et al., 1989) and the consensus
sequence has been divided into two contiguous regions, a glycine-
rich loop followed by an a helix. The glycine-rich loop interacts with
the polyphosphate chain of the nucleotide, and the a helix appears to



rich \
loop a-helix f strand
mmm r--- --




Fig. 2.1 Consensus sequence forming part of a nucleotide
binding site. Sequences are compared from four sources, AK:
adenylate kinase (Fry et al., 1988), Y: yeast Fi-ATPase P
subunit (Takeda et al. 1985), B: bovine heart Fi-ATPase B
subunit (Walker et al., 1982), R: rat liver Fi-ATPase P subunit
(Garboczi et al., 1988), E: E. coli F1-ATPase P subunit (Walker et
al., 1982). Filled boxes represented absolutely conserved
sequences. The open box represents a conserved motif
consisting of one hydrophilic residue followed by two
hydrophobic residues and another two hydrophilic residues.
The conserved histidine corresponds to H36 of adenylate
kinase, H211 of the yeast Fi-ATPase p subunit, H179 of the
bovine B subunit, H177 of the rat liver 0 subunit, and H170 of
the E. coli 0 subunit.

interact with the purine moiety of the nucleotide. The effects of
mutations in the consensus sequence have been studied in several
proteins including p21, adenylate kinase, and yeast and E. coli F1-
ATPase P subunits (Adari et al., 1988, Reinstein et al., 1988, Mueller,
1989a, 1989b, others in Table 1.2) Amino acid substitutions in this
region generally lead to a decrease in catalytic activity.
Duncan et al. (1986) have proposed a three dimensional model
for the nucleotide binding site of the F1-ATPase 3 subunit which is
based on secondary structure predictions from the P subunit amino
acid sequence. The pattern of a helices and 3 strands in the region
of the predicted nucleotide binding site is similar to that of
adenylate kinase. The authors have therefore used the high
resolution structures of adenylate kinase as a scaffolding for their
model. Both X-ray crystallography and NMR spectroscopy have
been employed in the structural mapping of adenylate kinase.
These approaches have yielded virtually identical three dimensional
structures of the native enzyme, but there has been controversy
over the positioning of ATP in its binding site (Pai et al., 1977, Fry et
al., 1985, Tian et. al, 1988, Tsai and Yan, 1991). One specific source
of dispute has been the role of a particular histidine residue (H36)
which happens to be conserved in F1 P subunits. Early physical
studies led to the conclusion that H36 was involved in the catalytic
mechanism (discussed in Tian et al., 1988). The crystal structure of
Pai et al. (1977) seemed to corroborate this conclusion, but the NMR
studies of Mildvan and coworkers contradicted it (Fry et al., 1985,

1987). The NMR model predicts that the adenine moiety of MgATP
interacts with a hydrophobic cleft composed of residues I28, V29,
H36, L37, and L91. Residues 128 and V29 are located near the C-
terminus of the a helix of the consensus sequence. A 3 turn occurs
immediately after the end of this helix and a P strand follows (Figs.
2.2 and 2.3). H36 and L37 are located within the P strand and come
into proximity with the helix. In this model H36 is not situated close
enough to the polyphosphate chain to be involved in catalysis. The
F1 model of Duncan et al. adopts this binding configuration for ATP
(Fig 2.3).
Tian et al. (1988) performed site-directed mutagenesis of
adenylate kinase at position 36 and evaluated the kinetic and
structural properties of the resulting enzymes. They found that the
changes in kinetic constants caused by the mutations were
insufficient to justify a role for H36 in the catalytic mechanism, but
hydrophobic interactions between H36 and the adenine group of
MgATP were not ruled out. The most striking characteristic of the
mutant enzymes was their instability during purification. As the
substitutions for H36 became less conservative the resulting proteins
became more unstable, the order of stability being:
H36>H36Q>H36N>H36G. The resistance of the purified enzymes to
unfolding in the presence of guanidine hydrochloride was measured,
and the same hierarchy of stability was observed. The authors
concluded that while there may be hydrophobic interactions between

S<-- Conserved
Isoleucine (36)

Fig. 2.2 The hydrophobic cleft of adenylate kinase. Redrawn
from a high resolution structure presented in Fry et al. (1987).
The structure shows the main chain atoms from G19 to T39.
The imidazole ring of H36 is included and is shown projecting
into the cleft. The a-carbons of 128 and H36 are shown. 128
and H36 are homologous to L203 and H211 of the yeast F1-
ATPase j subunit.


Fig. 2.3 A model of the E. coli F1-ATPase B subunit nucleotide
binding site. This model is redrawn from Duncan et al. (1986). The
positions of amino acids homolgous to yeast P subunit L203 and
H211 are shown, as is the predicted orientation of ATP in the binding
site. Cylinders represent a helices and ribbons represent B strands.

H36 and the adenine moiety of MgATP, the most important role of
H36 is in the structural stabilization of the enzyme.
Garbozci et al. (1988) have approached the study of the FI P
subunit nucleotide binding site through the use of a synthetic
peptide. These authors synthesized a 50 amino acid peptide
homologous to the rat liver mitochondrial ATPase P subunit and
studied its interaction with ATP. The sequence of the peptide
encompassed both the glycine-rich loop and the a helical region of
the proposed nucleotide binding site. When the peptide was added
to a buffered solution containing 1 mM ATP, a precipitate formed
which was composed of the peptide and ATP. Examination of the
sequence of the peptide led to the suggestion that a histidine residue
is involved in substrate binding by the interaction of its positive
charge with a negative charge on the nucleoside triphosphate chain.
The histidine in question is homologous to H36 of adenylate kinase.
This idea is inconsistent with the NMR model of adenylate kinase and
the kinetic results of Tian et al., but is consistent with the X-ray
model. Fry et al. (1988) have examined the structure of a 45 amino
acid peptide of adenylate kinase by 2-D NMR, FTIR, and circular
dichroism spectroscopy. This peptide binds ATP and is homologous
to the rat liver peptide described above. They find that the tertiary
structure of this peptide is essentially the same as the corresponding
segment of native adenylate kinase. Chuang et al. (1992) have
performed similar studies with the rat liver FI p subunit peptide.
They find that it has transient tertiary structure and does not

resemble the adenylate kinase peptide. These results suggest that
the tertiary structure of this segment of the F1 p subunit may be
unlike the corresponding segment of adenylate kinase. If that is the
case, then the conserved histidine may play a different role in F1-
ATPase than it does in adenylate kinase.
In the rat liver Fi-ATPase 1 subunit H177 is homologous to
adenylate kinase H36. Thomas et al. (1992) have recently mutated
rat liver pH177 to asparagine. A three-quarter length C-terminal P
subunit fragment containing this mutation was expressed in E. coli
and purified. This peptide binds the the fluorescent ATP analogue
TNP-ATP with the same affinity and stoichiometry that the purified
wild type full-length p subunit does. The H177N peptide was
assumed to have folded correctly based on its solubility in E. coli and
its resistance to E. coli proteases. Better evidence of correct folding
would strengthen the conclusions drawn from this data, but the
nucleotide binding results are convincing nonetheless. These data
strongly suggest that H177 is not involved in substrate binding.
The resolution of the role of the conserved histidine in F1-
ATPase P subunits is the subject of chapters 2 and 3 of this thesis.
This work requires a system in which this residue can be examined
in the context of the FOF1 complex. In this respect, the peptide
approaches of Garboczi et al. and Thomas et al. provide interesting
preliminary data, but are not sufficient for firm conclusions
concerning the role of this residue in the native enzyme. The yeast
Saccharomyces cerevisiae provides a system which greatly facilitates

the study of single residues within the entire FoF1-ATPase by site-
directed mutagenesis. The gene encoding the B subunit (ATP2) has
been cloned and sequenced (Saltzgaber-Muller et al., 1983, Takeda et
al., 1985), and yeast hosts which express no P subunit are available
(Takeda et al., 1985). The well known variety of yeast shuttle
vectors and selectable genetic markers allows flexibility in the design
of experiments. For example, many recombinant genes are
expressed in yeast from multicopy plasmids. In cases where gene
dosage is important, the use of a single copy vector containing a
fragment of a yeast chromosome is an option. Another advantage of
the yeast system is their capacity for anaerobic growth. This allows
the recovery of mutations which would be lethal in a constitutively
aerobic organism. Yeast and other fungi also provide the model
systems for studying the assembly of the mitochondrial FoFi-ATPase.
The kinetic characteristics of the yeast enzyme are well established
and the enzyme is easily prepared in large quantities in either
membrane bound or soluble form (Takeshige et al, 1976). The yeast
Saccharomyces cerevisiae has therefore been selected as the system
in which to study the role of the conserved histidine of the F1 P
subunit in this work.
In the yeast Fi-ATPase P subunit the conserved histidine
residue corresponds to H211. In this chapter, the role of H211 has
been investigated by site-specific mutagenesis and subsequent
analyses of the mitochondrial ATPases from the mutant yeasts.
H211 has been replaced with an acid (aspartate), a base (lysine), a

polar structural analog asparaginee), a hydrophobic residue
isoleucinee), and a residue lacking sidechain bulk (alanine). The
rationale for these substitutions is as follows. If a positive charge is
important to the function of position 211, then the substitution of
aspartic acid for histidine should be extremely deleterious. On the
other hand, the insertion of lysine would provide a positive charge at
higher pH values than histidine does. This might result in a
broadening of the catalytic pH optimum, or in a shift of this optimum
to a higher pH value. The substitution of asparagine should provide
an amide nitrogen in roughly the same spatial location as the
nitrogen at position number 1 of the histidine imidazole ring. This
substitution tests the role imidazole nitrogen 1. An alternative role
of the histidine might be to provide hydrophobicity, perhaps to
interact with the adenine moiety of ATP. If this is the case, then the
substitution of isoleucine might impair function the least. The
substitution of alanine for histidine merely tests the need for a large
side chain at position 211 of the P subunit.

In addition to the mutations at position 211, a single mutation
was engineered at position 203. This codon, normally encoding
leucine, was altered to phenylalanine. L203 of the yeast FI P subunit
is equivalent to 128 of adenylate kinase (Fry et al.,1986). This
residue is predicted to form part of the hydrophobic cleft thought to
interact with the adenine moiety of MgATP. In the high resolution
structures of adenylate kinase it is located on the opposite side of the
cleft from the conserved histidine (Figs. 2.2 and 2.3). This mutation

results in an increase in hydrophobic bulk and tests the size
restriction on the hydrophobic group at position 203.
Briefly, the experimental approach was as follows.
Oligonucleotide-directed site-specific mutagenesis was performed on
ATP2 to generate the mutations described above. The resulting
genes were then expressed under the control of their natural
promoter in the yeast host strain AVY4-1 The chromosomal copy of
ATP2 carried in AVY4-1 has been destroyed by insertion of a copy of
LEU2, and it expresses no 0 subunit of its own. The phenotypes of
the yeasts expressing the mutant 0 subunits were then characterized
and the assembly and stability of the mutant ATPases was assessed.
The results of this chapter show that H211 is critical for
maintenance of the conformation of the enzyme complex and that it
is not required for catalysis. Furthermore, it is suggested that
positions 203 and 211 have a functional, and possibly a structural,
interaction. This supports the nucleotide binding site model of
Duncan et al. (1986) which predicts that these residues are in
proximity to each other.

E. coli NM522 (supE, thi,A(lac-proAB), A hsd59r-,m-),{F', proAB,
lacIqZAM15}), and SCS-1 (F-, recAl, endA1, gyrA96, thi, hsdR17 (rk-,

mk+), supE44, relAl) were used to recover and propagate
recombinant or mutated plasmids. L-broth (LB) consisted of 0.1%
tryptone, 0.05% yeast extract, and 0.05% NaCI, and was supplemented

with 100 mg/l ampicillin for the selection of bacteria carrying
plasmids of interest.
Yeast strains SEY2102 and AVY4-1 were gifts from Michael
Douglas. SEY2102 was the source of wild type Fi-ATPase in this
study, its genotype is MAT a ,ura 3-52, leu 2-3, 112, his 4-519, gal 2,
suc2-sA9. AVY4-1 is derived from SEY2102 by insertion of LEU2
into ATP2 (atp2 :: LEU2). This strain served as the host for expression
of wild type and mutated forms of of the P subunit. Strains carrying
mutated versions of ATP2 are named for the particular mutation, for
example the strain carrying the mutation H211N is simply named
PH211N. AVY4-1 carrying wild type ATP2 on ppOK (see below) is
referred to as strain pOK. The wild type strain SEY2102 was
transformed with a plasmid carrying URA3 (pVTU103) to allow
growth under the same conditions as the mutant strains. For
simplicity this transformed strain was named SEY103. Yeast minimal
media contained 0.67% Difco yeast nitrogen base, 20 mg/L adenine,
complete amino acid supplementation, and 2% glucose. Ethanol and
glycerol at 2% each were substituted for glucose when assaying
aerobic respiration. Yeast were transformed with plasmids by the
alkaline cation method (Ito et al., 1983) or by spheroplasting (Hinnen
et al., 1978). Transformants were selected by uracil prototrophy.
Plasmid pTZ19R (Pharmacia) was modified as follows (Fig. 2.4.).
A 195 bp BglI-SphI fragment was removed, the resulting ends were
made blunt by treatment with T4 DNA polymerase and the

Remove So. I
nd Pvu I
from Ap / in IIs


mT Promor

Fig. 2.4 Construction of pMFi. Details of the procedure are given in

plasmid was reclosed. The PvuI and Scal sites in the p-lactamase
gene (Ap) were simultaneously removed by site directed
mutagenesis as described below. The PvuII site at position 418 was
destroyed by the insertion of a 12 bp double stranded
oligonucleotide. The resulting 2.7 kb plasmid contains an fl origin,
Ap (ampicillin resistance), a T7 promoter, a reverse sequencing
primer, a pBR322 replication origin, and no restriction sites which
are unique in the sequence of ATP2. ATP2 was inserted into this
plasmid as a 1.5 kb HindIII fragment, and the resulting 4.2 kb
plasmid, pMFP (Fig 2.4), was used as the substrate for site-specific

Plasmid ppOK-H was a gift from Michael Douglas. It is a multi-

copy yeast shuttle vector containing a 2.8 kb yeast genomic Eco RI-
HindIII fragment which carries ATP2 and about 1.3 kb of 5' flanking
sequence (Fig 2.5). All of the mutations constructed in this study
were subcloned into this vector on restriction fragments which
replaced the corresponding wild type restriction fragment. pVTU103
is a yeast shuttle vector containing URA3 (Vernet et al.,1987).
SEY2102 was transformed with this plasmid to allow growth on
minimal media lacking uracil. ppOK-H, its derivatives, and pVTU103
were maintained in yeast by selection for uracil prototrophy.

Oligonucleotide-directed Site-specific Mutagenesis
Site-specific mutants were constructed by a gapped
heteroduplex technique (Fig 2.6) (Morinaga et al., 1989). Mutations

Xba I

Fig. 2.5 Plasmid pPOK. The details of this plasmid are discussed in

were designed so that they always either removed or introduced a
restriction site, this allowed enrichment for the desired mutation. For
example, pMF3 was digested with restriction enzymes at sites on
either side of the region to be mutagenized so that a fragment of
about 500 bp or less was eliminated. The larger restriction fragment
was then separated by agarose gel electrophoresis and eluted from
the gel. Another sample of pMFp was digested with a single
restriction enzyme at a unique site outside of the 500 bp gap region.
This sample was treated with calf intestinal phosphatase and
purified by electrophoresis as above. Approximately 3 pmole of the
two DNA fragments were then placed together in a microfuge tube
with 50 pmole of mutagenic oligonucleotide. The solution was
brought to 10 gl by the addition of 2 pl of a 10X ligase buffer (1 M
NaC1, 65 mM Tris pH 7.5, 80 mM MgCl2 and 10 mM p-
mercaptoethanol) and the DNA was denatured by incubation for 3
minutes at 1000 C. Single strands were allowed to reanneal and form
gapped heteroduplexes during 30 minute incubations at 300 C and
then 40 C. The solution was then incubated at OOC for 10 minutes.
The reaction volume was brought to 20 il by the addition of dATP,
dGTP, dCTP, and dTTP to 2.5 mM, ATP to 3.5 mM, beta
mercaptoethanol to 8 mM, and 5 units each of Klenow fragment of
DNA polymerase 1 and T4 DNA ligase. The reaction mixture was
incubated at 11-160 C for 8 hours to allow polymerization and
ligation of the gapped region. The resulting plasmids were used to
transform E. coli NM522. Competent cells were prepared by the


Digest with U
treat ends


i ^

Digest with A&B
Gel-purify large

Transform E. coli
Screen for site 'C'
Sequence across
gapped region

Fig. 2.6 The gapped

heteroduplex technique of site-directed

method of Hanahan (1983). Usually about 200 colonies were picked
and used to inoculate 20 0.5 ml cultures of L-broth with 100 mg/L
ampicillin. Plasmid DNA was isolated from these heterogenous
cultures by boiling lysis (Maniatis et al., 1982) and enriched for the
presence or absence of a restriction site as follows. If a mutation
resulted in the introduction of a restriction site, the plasmid was
digested with an enzyme recognizing that site, separated from uncut
plasmid DNA on an agarose gel and purified. The DNA was then
ligated and passed through E. coli once again after which plasmids
from individual colonies were screened for the introduction of the
restriction site. DNA was isolated from positive clones by alkaline
lysis, and the region which had constituted the gap in the
heteroduplex was sequenced and then substituted as a cassette for
the corresponding region of a wild type clone carried on ppOK-H.

If the desired mutation resulted in the removal of a restriction
site, then the plasmid DNA from heterogenous culture was digested
with the enzyme which recognized the site targeted for removal, and
another enzyme for which the plasmid had a unique site. The
resulting molecules were separated by electrophoresis and once cut
plasmids were gel-purified, ligated, and used to transform E. coli.
Single colonies were then screened for the absence of the site in
question, those clones lacking it were sequenced, and positive clones
were inserted into ppOK-H as described above.

Nucleotide changes in codon 211 were as follows: CAT (H211)
was altered to AAT (H211N), GAT (H211D), AAG (H211K), ATT

(H2111), and GCT (H211A). These changes resulted in the elimination
of a unique Nco I site. The nucleotide change in codon 203 was TTG
(L203) to TTC (L203F). This resulted in the introduction of an Eco RI
site. Mutagenic oligonucleotides and sequencing primers were
synthesized and sequencing was performed by the core facilities of
the University of Florida Interdisciplinary Center for Biotechnology
Preparation of Submitochondrial Particles and Purification of ATPase
Yeast were grown in minimal media to an absorbance of 1.8 at
610 nm, collected by centrifugation at 3,000 X g for 5 minutes at 40
C, and resuspended in chilled breaking buffer (0.6 M sucrose, 20 mM
Tris pH 7.5, 1 mM EDTA, 1 mM PMSF, 1 mM para-
aminobenzamidine). All subsequent steps were carried out at 0-40 C.
Yeast were broken in a Bead Beater (Bio-Spec products, Bartlesville,
OK) using five breaking cycles of two minutes separated by 6-10
minutes cooling in ice water. Cellular debris was removed by
centrifugation 2-4 times at 3000 X g and crude mitochondria were
collected by centrifugation for 20 minutes at 20000 X g. The pellet
was resuspended in breaking buffer at 5-20 mg protein/ml buffer,
and submitochondrial particles were generated by sonic oscillation
for 1 minute in a Heat Systems W-225 sonicator. Debris was
removed by centrifugation at 20000 X g for 10 minutes.
Submitochondrial particles were collected by centrifugation at
100,000 X g for 90 minutes, resuspended in breaking buffer at 5 mg
protein/ml, and stored at -750 C.

F1-ATPase was purified by chloroform extraction and DEAE
anion exchange chromotography essentially as described (Mueller,
1988) except that gel filtration was omitted. Submitochondrial
particles (10-20 mg protein/ml) were extracted with 0.5 volume of
chloroform. ATP was added to 2 mM and debris was removed by
centrifugation for 10 minutes at 2000 X g at room temperature. The
aqueous phase was transferred to a 50 ml Corex tube and
centrifuged at 48000 X g for 30 minutes at room temperature. The
pH of the supernatant was adjusted to 7.5 with 0.1 N NaOH and was
loaded onto a 10 X 2.5 cm column of Whatman DE-52 anion exchange
resin which had been equilibrated with breaking buffer plus 2 mM
ATP (buffer B). The column was washed with 100 ml of buffer B,
and then washed with 100 ml of buffer B plus 20 mM K2S04. FI was
eluted with Buffer B plus 75 mM K2S04. The fractions containing 80%
of the activity were pooled, ammonium sulfate was added to 72% of
saturation and the precipitate was stored at 40 C.
FoF1-ATPase was purified by sodium cholate/n-octyl P-D-
glucopyranoside extraction (Rott and Nelson, 1981). Crude
mitochondria were prepared as described above. Sodium cholate and
n-octyl P-D-glucopyranoside were added to 0.5% and 1% respectively
and the mixture was incubated for 20 minutes at 00 C. The
suspension was centrifuged at 200000 X g for 30 minutes.
Ammonium acetate was added to the supernatant to 37% saturation
and the suspension was incubated at 00 C for 20 minutes. The
precipitate was removed by centrifugation at 10000 X g for 10

minutes, and ammonium sulfate was added to the supernatant to
48% saturation. The suspension was incubated at 00 C for 20 minutes
and precipitated material was collected by centrifugation at 10000 X
g for 10 minutes. The pellet was resuspended at a protein
concentration of 5-10 mg/ml in 30 mM Tris succinate (pH 6.5), 0.2%
Triton, 0.1 mM ATP, 0.5 mM EDTA, and 0.1% acetone washed L-a-
phosphatidylcholine vesicles (added as 40 mg vesicles/ml 80 mM
Tricine pH 8.0). Vesicles bearing reconstituted FoFI were purified by
centrifugation on 7 to 30% sucrose gradients at 100000 X g for 24
hours at 40 C. Gradients were fractionated and ATPase activity
appeared at 4-6 ml from the bottom of 12 ml gradients.
Assay of ATPase Activity and Determination of Steady State Kinetic
ATP hydrolysis was measured at 300 C by a coupled assay (Ebel
and Lardy, 1975) in a reaction cocktail containing 20 mM Tricine pH
8.0, 0.45 mM NADH, 2 mM PEP, 6.5 mM KCI, 2 mM MgCl2, and 2
units/ml each of lactate dehydrogenase and pyruvate kinase. One
control reaction containing the reaction cocktail and
submitochondrial particles without ATP was run with every five
experimental reactions. The rate in this control was subtracted from
the experimental reactions to eliminate any rate contributed by
direct oxidation of NADH by constituents of the submitochondrial

Immunologic Characterization of Submitochondrial Particles
Submitochondrial particles were separated on 10% SDS-
polyacrylamide gels (Laemmli, 1971) the F1-ATPase a and/or p
subunits were detected by western blot as described (Maniatis et al,
1989). Anitsera raised against the a subunit, the P subunit, and the
entire yeast FI complex were provided by Dr. Alfred S. Lewin.

Phenotypic Characterization of Mutant Yeasts
Six mutations were engineered into ATP2 as described in
Methods. These mutations were: H211N, H211D, H211K, H2111,
H211A, and L203F. Sequenced restriction fragments containing the
mutations were used to replace the corresponding restriction
fragment in ppOK. Plasmid ppOK is a multicopy yeast shuttle vector
which carries all the necessary information for the transcription of
ATP2 under the control of its natural promoter. The resulting
plasmids were named for the mutation they carried: pN211, pD211,
pK211, pI211, pA211, and pF203. The yeast strain AVY4-1
(atp2::LEU2 ) was transformed with each of these plasmids
separately. This resulted in six mutant strains, each expressing a
different mutated 3 subunit. The ability of the mutant strains to
respire aerobically was assayed by observing the size of colonies
formed on a nonfermentable substrate (Fig. 2.7). Of the six mutant
strains, only PL203F and pH211N were capable of aerobic respiration.
The other strains formed no colonies. The growth rate of pL203F was
greater than 3H211N but less than that of the wild type. Because




k n ..

> 0 z
b" 0 .5
s' i"

3 c o
a Vv

m ny ;S te

these strains were capable of aerobic respiration, it was concluded
that neither L203 nor H211 is required for catalysis by the wild type
FI-ATPase. Plasmid pPOK was used to transform AVY4-1 and the
resulting strain was phenotypically wild type.
To assess whether the position 211 mutations were dominant
or recessive, partial polyploids were constructed by transformation
of SEY2102 (ATP2) with each of the shuttle vectors containing the
position 211 mutations. SEY2102 is the parent strain of AVY4-1.
The resulting strains contained a single copy of ATP2 and multiple
copies of atp2 genes, and were therefore polyploid for the ATP2
locus. These yeasts were then supplied with a nonfermentable
carbon source and their growth was observed (Fig 2.8). The growth
rate of each of the resulting strains was indistinguishable from the
wild type, therefore each position 211 mutation is phenotypically
Five plasmids were constructed which carried doubly mutated
atp2 genes. The L203F mutation was paired with each of the
position 211 mutations, so that each plasmid carried one copy of atp2
which was mutated at two sites, positions 203 and 211. The
resulting plasmids were named pL203F/H211N, pL203F/H211D etc.
These plasmids were used to transform AVY4-1 and the capacity of
the resulting strains to respire aerobically was assessed as above.
Each of the doubly mutated strains was capable of aerobic
respiration (Fig 2.9). L203F is therefore an intragenic suppressor of
mutations H211D, H211K, H2111, and H211A. Kinetic data will be

6 2

5 3


Fig. 2.8 The position 211 mutations are recessive to wildtype.
SEY2102 was transformed with each of five shuttle vectors carrying
position 211 mutations. Single colonies were picked and grown as
described in figure 2.7. Plates were incubated at 30o C for 3-5 days.
1: SEY103, 2: SEY/PH211D, 3: SEY/BH211I, 4: SEY/pH211A, 5:
SEY/PH211N, 6: SEY/PH211K.

* .r-

- .U

0.2 0
S 3 .
i 4 C O

Se ,o ,

sg g-

U. BW Id u m u

presented in the next chapter to show that L203F also suppresses
H211N. These data are suggestive of a functional, and possibly
structural, interaction between L203 and H211 of the P subunit.
Characterization of the Stability and Structure of the Mutant ATPases
In order to evaluate further the effects of mutations in the
yeast F1-ATPase P subunit, it was necessary to characterize the
mutant ATPases with respect to their assembly, stability, and
structure. Evaluation of several characteristics was required, these
included subunit composition, stability of the enzyme over time, and
stability under conditions of assay. The first step in these analyses is
the purification of the enzymes.
Stability of chloroform extracts
Submitochondrial particles were prepared from each of the
singly mutated strains, and each preparation showed oligomycin
sensitive ATPase activity (see below). Chloroform extracts of the
submitochondrial particles were then subjected to DEAE anion
exchange chromatography as described in Methods. No significant
ATPase activity was detected in any fraction of the eluates. The
ATPase activities of the chloroform extracts were assayed before
chromatography and were found to be unstable. The addition of 20%
glycerol to submitochondrial particles prior to chloroform extraction
did not stabilize the activities of the extracts. The addition of 20%
glycerol to the ATPase assay likewise had no stabilizing effect. For
each mutant strain tested, the specific activities of the chloroform
extracts were greater than those of the submitochondrial particles,

but the rates of ATP hydrolysis decreased sharply after about 45
seconds of assay (Table 2.1). Specific activities of chloroform extracts
from 2.25 minutes to 5 minutes of assay were reduced by 10-fold to
14-fold relative to values from the first 45 seconds. Addition of ADP
to the reaction mixtures after 5 minutes resulted in a rapid loss of
absorbance, thus showing that the observed decreases in rates were
not due to oxidation of all the NADH in the reaction. These results
were reproducible in separate experiments using material from the
same extract. This is consistent with turnover-dependent, but not
time-dependent, dissociation of F1. The activities of the chloroform
extracts were stable for at least 20 minutes, but were not stable
enough to be purified by anion exchange chromatography which
takes about 7 hours.
Stability of sodium cholate and octyl glucoside extracts
The sodium cholate and octyl-G-glucopyranoside extraction
method of Rott and Nelson (1981) was used to reconstitute active
wild type FoFI complexes in phospholipid vesicles. Preparations of
wild type ATPase reproducibly showed oligomycin-sensitive ATPase
activity at 4-6 ml from the bottom of a 12 ml sucrose gradient. This
is considered to be the gentlest purification procedure for
mitochondrial ATPases. Nonetheless no ATPase activity was
recovered from the mitochondria of any strain carrying substitutions
at positions 211 or 203 using this protocol. These results suggest
that mutations at position 211 destabilize the F1 complex, as does the
substitution of phenylalanine for leucine at position 203.

Table 2.1 Specific activities of chloroform extracted ATPases
from mutant yeast strains.

Specific Activity
Strain 0-0.75 min. 2.25-5 min.
PH211D 10.9 1.1
PH211I 6.9 0.5
PH211A 13.5 1.3
PjH211N 8.4 0.9
PH211 K 5.2 0.4

Submitochondrial particles were adjusted to a pH of 6.25
extracted with 0.5 volumes of chloroform, and the ATPase
activities of the extracts were determined by coupled assay as
described in Methods. The concentration of ATP in the assay
was 12 mM and the pH of the assay medium was 6.25. Specific
activity is expressed in gmol X min-lmg-1 protein. Values
shown for mutants are the average of three experiments.
Wildtype values were in the range of 15-20 gmol X min-lmg-1
and did not decrease over the course of the assay.

Submitochondrial particles prepared from the mutant strains retain
most of their ATPase activity after months of storage at -750 C, but
separation of the FI portion from the membrane results in
destabilization of the complexes. Association with membrane bound
FO appears necessary to stabilize these enzymes.
Because the ATPases from the mutant strains could not be
purified, all of the kinetic and physical characterizations in this study
were performed on submitochondrial particles. Unless otherwise
indicated, all statements regarding ATPase activities of the mutants
therefore refer to activities of submitochondrial particles isolated
from the mutant strains. For purposes of comparison,
submitochondrial particles from the wild type were used as well,
rather than pure Fi. The stability over time of the membrane bound
ATPases was assayed as follows. Submitochondrial particles were
incubated at 300 C and their ATPase activity at pH 8.0 was measured
several times over 4 hours. Strains SEY103, POK, and PH211I showed
no loss of ATPase activity in this experiment, while strains pH211D,
PH211A, and PH211K displayed 84%, 71%, 70%, and 67% of their
original activities after 4 hours at 300 C. ATPase activities during
each five minute assay did not decrease during the assay period.
This experiment was repeated with incubation of the
submitochondrial particles at 00 C rather than 300 C, and no loss of
activity was seen in any strain after 3 hours of incubation. In all
experiments in this study in which ATPase activity is measured, the

submitochondrial particles were kept on ice for less than three hours
prior to assay.
Immunological characterization of the mutant ATPases
In order to assess the subunit composition of the mutant
ATPases a series of immunoblots were performed. Submitochondrial
particles from wild type and mutant yeast were separated by SDS-
PAGE, transferred to PVDF membranes and probed with antisera
raised against either purifed FI a subunits or purified Fl. The results
of these experiments can be summarized as follows. The quantity of
a subunit present in submitochondrial particles from host strain
AVY4-1 was influenced by the presence of the 3 subunit. For
example, particles from strain POK contained as much a subunit per
unit protein as did particles from SEY103 (Fig. 2.10). The quantity of
a subunit in particles from the mutant strains varied with the
preparation, but was never more than about one third the amount
seen in POK or SEY103 (Fig. 2.11). The quantity of a subunit in
AVY4-1 particles, which contained no detectable P subunit, was at
least 10-fold less than in SEY103 or POK particles (Fig. 2.12). The
ratio of P to a subunits in the mutant strains and in POK was slightly
higher than that of SEY103 (Fig. 2.10). This may reflect the presence
of unassembled 1 subunits due to the presence of multiple copies of
the expression plasmid and subsequent overexpression. The
quantities of a subunits detected in these assays allow an upper limit
estimate of the amount of ATPase complexes which can be formed in
the mutant strains under the growth conditions described.


S- o
C_.. in_. C _. n_ _0W U F1

a subunit

B subunit

Fig. 2.10 Quantitation of a and P subunits in submitochondrial
particles from mutant and wildtype yeasts and soluble yeast
Fl-ATPase. Submitochondrial particles were solublized in SDS-
PAGE sample buffer, and immunoblots were performed as
described in Methods using antiserum raised against yeast F1-
ATPase. Lanes PK211-SEY103: 30 gg protein loaded per lane.
Lane Fl: 3 gg of pure yeast Fi-ATPase from strain SEY2102

15 30 45 60
1 1 1 1


5 10 15 30
1 I I 1

5 +- a subunit


SEY103 PH211A

30 15 10 5 60 45 30 15


se- a subunit

Fig. 2.11 Quantitation of the a subunit in submitochondrial
particles from mutant and wildtype yeast. Submitochondrial
particles were solublized in SDS-PAGE sample buffer and
immunoblots were performed as described in Methods.
Antiserum was specific for the a subunit. The amount of
protein loaded in each lane is indicated in gg.

AVY4-1 SEY103

20 10 5 20 10 5

a subunit

B subunit

Fig. 2.12 Relative amounts of a subunit in yeast strains AVY4-
1 and SEY103. Immunoblots of submitochondrial particle
protein were performed as described in Methods using
antiserum raised against yeast Fi-ATPase. The amount of
protein loaded per lane is indicated in gg. The migration of the
a and B subunits is shown.

The total ATPase complexes formed per unit mitochondrial protein in
the strains carrying mutations at position 211 cannot be greater than
one third of that in strains pOK and SEY103. The minor subunits were
not detectable by the anti-Fi antiserum in the widtype or mutant
submitochondrial particles.
Oligomycin sensitivity of the mutant ATPases
The oligomycin sensitivity of the mutant ATPases was
evaluated as an indirect test of ATPase assembly. Inhibition of
mutant ATPases by oligomycin requires that oligomycin sensitivity
conferring protein (OSCP) is assembled into the complex (Tzagoloff,
1970). Fig 2.13 shows the effects of oligomycin and methanol on
ATP hydrolysis by submitochondrial particles from the wild type and
mutant yeasts. Oligomycin was added to the reaction as a methanolic
solution from a 2 mg/ml stock, and equivalent amounts of pure
methanol were added to parallel reactions as a control. ATPase
activity was measured at a substrate concentration of 6 mM. ATPase
activities of the wild type and mutant strains were not significantly
affected by methanol concentrations up to 2%. Maximal inhibition of
the ATPases occurred at an oligomycin concentration of 20 Vg/ml
(this corresponds to a methanol concentration of 1%). At pH 8.0
ATPase activities of the mutants were inhibited by 30-50%, while
SEY103 and pOK strains were inhibited by 93.3% and 93.8%
respectively. The experiment was repeated at pH 6.5 with little
change in the extent of inhibition of the mutant ATPases, while
inhibition of POK by 20 tg/ml oligomycin decreased to 64%. The








0 1 2 3MeO

4 5

0 20 40 60 80 100
ug ollgo/ml

Fig. 2.13 The sensitivity of wild type and mutant Fl-ATPases
to oligomycin. Panel A: the effect of methanol on ATP
hydrolysis. Panel B: the effect of oligomycin on ATP hydrolysis.
Oligomycin was added to the concentrations shown from a 2
mg/ml methanolic stock solution. The amount of methanol in a
reaction containing 20 gg/ml oligomycin is 1% v/v. Rates of
ATP hydrolysis were measured as described in Methods and
are expressed as percentages of the rates measured in the
absence of methanol or oligomycin. Rates are averages of 2-3
measurements Open circles: SEY103, closed circles: PH211D,
open squares: PH211I, closed squares: pH211A, open triangles:
PH211N, closed triangles: PH211K, X: BOK.

n ... .... ,

reduced sensitivity to oligomycin of the ATPases from the mutant
strains is suggestive of structural perturbation in these complexes,
however it must be concluded that the oligomycin sensitivity
conferral protein is assembled into each complex.
The effect of organic solvents on the mutant ATPases
In the course of measuring oligomycin sensitivity, it was noted
that a methanol concentration of 5% caused a slight decrease in the
rate of ATP hydrolysis by the mutant enzymes. This inhibition
increased when the methanol concentration was adjusted to 10%. It
was proposed that this apparent inhibitory effect was due to the
destabilization of the enzyme complexes by the increasing nonpolar
character of the solvent. In order to assess further the stability of
the mutant enzymes, their activities in the presence of several
organic solvents was investigated. We chose to use 2-propanol,
ethanol, methanol, and dimethyl sulfoxide on the basis of their
varying degrees of polarity. Figs. 2.14-2.17 show the effects of these
solvents on the activities of the wild type and mutant ATPases. The
general trend for the effect of 2-propanol is similar to that for
ethanol, although 2-propanol does not seem to have as negative an
effect at 1% and 2%, and in the cases of 3H211A and PH211K it is
stimulatory at these concentrations. Methanol had little effect on
ATP hydrolysis by the mutants or the wild type. Dimethyl sulfoxide
stimulated the activities of the mutant ATPases but inhibited the
wild type at concentrations of 5% and 10%. The stimulatory effect of




o 60

0 40


0 2 4 6 8 10 12

% 2-Propanol (v/v)

Fig. 2.14 The effect of 2-propanol on ATP hydrolysis by
mutant and wild type ATPases. Rates of ATP hydrolysis in the
presence of 2-propanol are expressed as percentages of the
rates determined in the absence of ethanol and are averages of
2 or 3 determinations. ATP hydrolysis was measured by
coupled assay as described in Methods. Reactions included 12
mM ATP and 150 ugg submitochondrial particles/ml, except for
SEY103 for which 20 gg/ml were used. Open circles: SEY103,
closed circles: pH211D, open squares: 3H211I, closed squares:
PH211A, open triangles: pH211N, closed triangles: pH211K, X:


0 60-

o 40-


0 2 4 6 8 10 12

% Ethanol (v/v)

Fig. 2.15 The effect of ethanol on ATP hydrolysis by mutant
and wild type ATPases. The experiment was performed
precisely as described in Fig. 2.14. Rates of ATP hydrolysis in
the presence of ethanol are expressed as percentages of the
rates determined in the absence of ethanol and are averages of
2 or 3 determinations. Open circles: SEY103, closed circles:
pH211D, open squares: PH211I, closed squares: PH211A, open
triangles: PH211N, closed triangles: pH211K, X: PL203F.

o 80-

o 40


0 1-
0 2 4 6 8 10 12

% Methanol (v/v)

Fig. 2.16 The effect of methanol on ATP hydrolysis by mutant
and wild type ATPases. The experiment was performed
precisely as described in Fig. 2.14. Rates of ATP hydrolysis in
the presence of methanol are expressed as percentages of the
rates determined in the absence of ethanol and are averages of
2 or 3 determinations. Open circles: SEY103, closed circles:
PH211D, open squares: pH211I, open triangles: pH211N, X:

S 100-

o 80



0 1 1 -
0 2 4 6 8 10 12

% Dimethylsulfoxide (v/v)

Fig. 2.17 The effect of dimethylsulfoxide on ATP hydrolysis by
mutant and wild type ATPases. The experiment was
performed precisely as described in Fig. 2.14. Rates of ATP
hydrolysis in the presence of dimethylsulfoxide are expressed
as percentages of the rates determined in the absence of
dimethylsulfoxide and are averages of 2 or 3 determinations.
Open circles: SEY103, closed circles: PH211D, open squares:
PH211I, closed squares: pH211A, open triangles: pH211N, closed
triangles: PH211K, X: BL203F.

dimethyl sulfoxide on the mutants was greatest at 5% and 10%. The
effect reached a plateau for PH211A, pH211I, PH211I, and 1L203F at
5%. Stimulation decreased at 10% for 3H211N and P3H211D.
The roles of positions L203 and H211 of the yeast FI ATPase 1
subunit were investigated by substituting Phe for Leu at position
203 and Asn, Lys, Asp, Ile, and Ala for His at position 211.
Mutagenesis was performed in a plasmid which was specifically
modified to facilitate mutagenesis of ATP2. Sequenced restriction
fragments containing the mutations were used as cassettes to replace
the corresponding restriction fragments in wild type copies of ATP2
carried on the yeast shuttle vector pPOK-H. The resulting plasmids
were used to transform yeast strains AVY4-1 (atp2::LEU2) and
SEY2102 (ATP2).
H211 and L203 are not Required for Aerobic Respiration or Catalysis
Of the five amino acids substituted for histidine, asparagine is
the best structural mimic. The results in Figure 2.7 show that strain
PH211N is the only position 211 mutant capable of aerobic
respiration. This suggests that a positive charge on H211 is not
involved in the wild type catalytic mechanism. The asparaginyl
amide nitrogen of H211N may occupy nearly the same spatial
location as the number 1 nitrogen of the imidazole ring of H211. It is
possible that position 211 may be involved in hydrogen bonding, and
that a histidine at this residue is better positioned to achieve this
than asparagine. A hydrogen bond at position 211 might be required

for either assembly or structural stabilization. None of the other
position 211 mutants respires aerobically, suggesting that either the
sidechain shape or the hydrogen bonding potential of this residue is
critical for enzyme function. Tian et al. (1988) have performed site-
directed mutagenesis on the corresponding histidine of adenylate
kinase (H36). As discussed above, the primary result of the amino
acid substitutions was the destabilization of the enzyme. The authors
offered three explanations for this based on the three dimensional
models available. They proposed that H36 either hydrogen bonded
with C25, formed a charge-transfer complex with C25 and D93, or
protected C25 by shielding it while also forming a hydrogen bond
with D93. In each case the number 3 nitrogen of the imidazole group
is predicted to be involved in an electrostatic interaction. In this
context a substitution of glutamine for H211 may have been more
instructive than the substitution of asparagine for H211. In any case,
the results suggest that the shape of the residue at position 211 is
more important than its hydrophobic character, and the ability to
form a hydrogen bond may also be important.
H211 and L203 are Required for the Proper Assembly of the Fi-
ATPase Complex
Phenotypic analyses
Before meaningful kinetic analyses can be performed on the
mutant enzymes, the issue of assembly of the FoF1 complex in the
mutant strains must be addressed. Much of the data presented in
this chapter is pertinent to this point. The recessive nature of the

position 211 mutations was demonstrated by transforming yeast
strain SEY2102 (ATP2) with multi-copy yeast shuttle vectors
carrying atp2 mutated at codon 211. This resulted in yeast which
were polyploid for the ATP2 locus, but which carried only one wild
type allele. The fact that these yeast were phenotypically wild type
raises the question of whether or not the mutant P subunits assemble
with wild type subunits to form hybrid Fl complexes. Noumi et al
(1987) and Rao and Senior(1987) reported the formation of E. coli Fl
hybrid complexes using mutated 1 and a subunits respectively. Rao
and Senior generated uncA- E. coli strains incapable of aerobic
respiration. They then purified F1, dissociated it, and reassociated
the complexes in the presence of varying amounts of wild type a
subunit. The reassembled complexes containing both wild type and
mutated a subunits had diminished catalytic activity relative to the
wild type. Noumi et al. reported similar results with mutated p
subunits. The formation of hybrid complexes in the yeast system
should also result in a decreased catalytic capacity which might be
manifest as a decreased growth rate on a nonfermentable carbon
source. Based on growth rates for the polyploid yeast strains in this
study, it seems likely that the mutated P subunits are preferentially
excluded from the FI complex in favor of the available wild type
subunits. This might be expected if the mutated P subunits were not
refolded correctly after transport into the mitochondrial matrix. It is
possible that the mutations affect the interaction of the P subunits
with hsp60MIF4 or with the a subunit prior to assembly. This raises

the questions of how well the mutant p subunits are assembled into
the FoF1 complex in the absence of wild type P subunits, and what
pleiotropic effects do the mutations have on the assembly of the
other subunits?
Immunologic characterization
Because the mutant ATPases could not be purified, analysis of
their subunit compositions is technically difficult. Western blot
analyses of submitochondrial particles from the mutant strains can
show the identities and relative quantities of subunits present in the
mitochondria, but may yield misleading information regarding
assembly. Immunoprecipitation of the complexes with antisera
directed to a single subunit may also be misleading because of the
instability of the mutant complexes. With these problems in mind,
western blot analyses using polyclonal antisera directed against
either the a subunit or the whole Fl complex were carried out on
submitochondrial particles from the mutant and wild type strains.
The results show that the mutations at position 211 affect the
assembly of the FI complex. The nature of this effect is not known
beyond the fact that the mutant strains assemble no more than one
third as much FI as the wild type. It seems likely that the rate of
either assembly or disassembly or both has been affected such that
the equilibrium amount of complexes formed is less than in the wild
type. This would be consistent with the exclusion of the mutated
subunits from the ATPases formed by the heterozygous partial
polyploid strains. The issue of subunit composition of the mutant

ATPases with respect to the minor subunits was not resolved by
immunological analysis and will be addressed below.
The observed influence of the P subunit on the amount of a
subunit detected in submitochondrial particles is in apparent
contradiction to the results of Takeda et al. (1985) who found that
the absence of the 13 subunit had no effect on the import of the a
subunit into the mitochondria of MDY2102 (an atp2::Leu2 strain
derived from SEY2102). These authors immunoprecipitated a and P
subunits from detergent extracts of whole mitochondria. The
discrepancy in results may be due to the use of submitochondrial
particles in this study, rather than whole mitochondria. It is possible
that after disruption of the mitochondria, unassembled a subunits

were released and not recovered in the submitochondrial particle
Assembly of minor subunits
The oligomycin sensitivity conferral protein forms part of the
stalk region between Fj and Fo. Uh et al. (1990) have shown that the
presence of OSCP is required for the binding of F1 to Fo. That this
protein assembles into the mutant ATPase complexes is obvious
because they are each sensitive to oligomycin. However, the degree
of sensitivity to oligomycin is much reduced relative to the wild type.
This is difficult to interpret unambiguously, but is suggestive of
structural perturbation due to the P subunit mutations. The
assembly of the E subunit into the complexes can be inferred
indirectly since it has been shown (Guelin et al., 1993) that e null

mutants have no oligomycin sensitive ATPase activity. The presence
of the y and 8 subunits has not been demonstrated. In the E. coli
system, all eight subunits of FoFi are required for ATP synthase
activity (Senior 1990). This fact provides some support for the
contention that the FOF1 moiety of strain pH211N is fully assembled.
An argument for the full assembly of the mutant F1 complexes based
on kinetic data will be presented in the next chapter.
The Effect of Mutations at Positions 211 and 203 on the Stability of
the FI Complex
Several lines of evidence suggest that H211 and L203 of the P
subunit are important for the structural integrity of the F1 complex.
Attempts to purify ATPases from the mutant strains using
techniques which involved either permanent or temporary removal
of the F1 portion of the enzyme from membrane bound FO failed. On
the other hand, activities of submitochondrial particles from the
mutant strains were stable after months of storage at -750C. This
suggests that alteration of either position 211 or 203 destabilizes the
F1 complex in such a way that it dissociates upon removal from Fo.
Measurement of the activities of chloroform extracts of
submitochondrial particles indicated that dissociation of the ATPases
is either turnover dependent or is at least facilitated by catalysis
(Table 2). Dissociation probably also occurs over time, but the
activities of the chloroform extracts were stable for at least twenty
minutes in the absence of MgATP.

The instability of the mutant ATPases was also demonstrated
by the effects of ethanol and 2-propanol on catalysis by
submitochondrial particles (Figs. 2.15 and 2.16). It was proposed that
if the mutant ATPases were not folded entirely correctly, then
exposure to organic solvents might cause further structural
perturbation resulting in decreased activity. Ethanol and 2-propanol
inhibited ATP hydrolysis by the mutant enzymes at concentrations
which had little effect on the wild type ATPase. Inhibition increased
with the concentration of ethanol or 2-propanol in the assay.
Methanol had little effect on catalysis by ATPases from mutant or
wild type strains at the concentrations studied. This is consistent
with inhibition being due to increased nonpolarity of the assay
Taken together, the reduced sensitivity of the mutant ATPases
to oligomycin, the inhibitory effects of ethanol and 2-propanol, and
the apparent instability of the altered F1 complexes when separated
from the membrane, provide strong evidence that H211 is important
for the stabilization of the FI complex. Replacement of histidine at
position 211 with amino acids of varying characteristics i.e. an acid, a
base, hydrophobic residues, or a polar isosteric replacement results
in destabilization of the complex. Replacement of leucine at position
203 with phenylalanine apparently causes similar effects.
The identification of L203F as an intragenic suppressor of the
position 211 mutations was an unexpected result which bears
directly on the question of the structural and functional interaction

of L203 and H211. There are several instances in the study of FoFI-
ATPase alone in which intragenic or extragenic supressor mutations
have been interpreted as evidence of structural or functional
interaction between the positions at which the mutations occur
(Kumamoto and Simoni 1986, 1987, Cain and Simoni, 1988,
Nakamoto et al., 1993). As discussed above, high resolution
structures of adenylate kinase have been used as a framework for
proposing a structure for the E. coli FI-ATPase P subunit nucleotide
binding site (Duncan et al.,1986). In such a model L203 would reside
near the end of an a helix which follows a glycine rich loop. After
the C-terminus of the helix there is a hairpin turn in the peptide
backbone which gives rise to the proposed hydrophobic cleft with
which the adenine or ribose moiety is associated. In this model L203
is situated on the opposite side of the cleft from H211 (Figs. 2.2 and
2.3). It is conceivable that the sidechains from positions 203 and
211 interact and that substitution for either of them distorts the
conformation of the complex in a similar manner. Each of the amino
acids which were substituted at position 211 takes up less volume
than histidine. On the other hand, phenylalanine is bulkier than the
leucine at position 203. One simple explanation for the suppression
of the position 211 mutations by L203F is that positions 203 and 211
interact physically. When the side chain bulk at position 211 is
lessened by substitution of a smaller amino acid, the interaction does
not occur. An increase in side chain bulk at position 203 then

reestablishes the interaction but not to the same extent as in the wild
Another speculative interpretation would be that the side
chains at position 203 and 211 do not interact physically, but instead
share a functional role. For example, both sidechains might interact
with the adenine moiety of ATP as it associates with the hydrophobic
cleft. Substitution of a smaller amino acid for histidine at position
211 could decrease the interaction of this residue with adenine.
Insertion of a phenylalanine at position 203 might alleviate the
decrease in hydrophobic character of the cleft, or it might simply
"push" the adenine closer to position 211. A similar scenario could
be envisioned if the role of H211 was to stabilize the nucleotide
binding site by hydrogen bonding with the another residue
analogous to D93 of adenylate kinase. Removal of the hydrogen
bonding group at position 211 might interfere with the stabilization
of the nucleotide binding site. Phenylalanine at position 203 might
partially suppress the binding defect by increasing the hydrophobic
nature of the cleft. This idea may be the most consistent with the
results in that the H211N is the least deleterious mutation. If
phenylalanine at position 203 alters the local structure of the protein
so that the asparaginyl amide nitrogen can reestablish a stabilizing
hydrogen bond, the original defect might be suppressed.
Alternatively, the suppression might be explained by a
mechanism which does not require the interaction of residues 203
and 211. The L203F mutation may alter the structure of the P

subunit in such a way that further mutations at position 211 have
little effect on the function of the 0 subunit. This can be envisioned

as follows. While H211 is not required for catalysis, it may help
maintain the structure of the active site. Single mutations at position
211 may result in interference with the catalytic mechanism,
resulting in the phenotype of the position 211 mutants. A situation
can be imagined where the mutation at position 203 alters the local
structure of the protein such that position 211 is removed from its
normal space in the tertiary structure. Position 211 would then be
unavailable to perform its stabilization role. This results in the
phenotype of L203F. Suppression of the position 211 mutations may
result from their removal by the mutation at position 203. Position
211 would not fulfill its role in structural stabilization, but neither
would it cause interference with the catalytic mechanism. A
phenotype similar to the L203F mutant would then be expected.
Each of these explanations is highly speculative and,
unfortunately, difficult to test. What is clear from the results of this
chapter is that H211 is not required for the chemistry of catalysis in
the yeast F1-ATPase. The same can be said of L203. H211 and L203
are each important for the structural stability of the enzyme as
evidenced by the failure to purify enzymes with mutations at these
residues. Mutations at H211 also have a negative effect on the
assembly of the ATPase complex. This is probably due to a role in
the folding of the 1 subunit. The data are also consistent with a

functional and/or structural interaction between L203 and H211.

This would in turn be consistent with the three dimensional model
for the nucleotide binding site proposed by Duncan et al. One further
interesting note is the report of Rao et al. (1988) which showed that
the E. coli Fi-ATPase could not be reassembled from pure subunits
unless the P subunits had bound ATP. If the mutated 1 subunits do
not fold correctly after transport into the mitochondria, then they
might bind ATP poorly and assemble into the complex more slowly
than usual. This would be consistent with the relatively low amounts
of ATPase detected in the mutant strains.
Although the mutations at position 211 destabilize the Ft
complex, the membrane-bound enzymes appear to be stable in the
absence of organic solvents. Submitochondrial particles from each of
the mutant strains retain most of their activity after months of
storage at -750 C, and incubation for several hours at 00 or 300 C
results in little or no loss of activity. The stability of the membrane-
bound ATPases provides the opportunity to characterize them
kinetically. Kinetic studies may yield information concerning the
possible roles of H211 in substrate binding, structural stabilization,
and any indirect contribution to the catalytic mechanism.

The results of chapter 2 show that H211 of the yeast F1-ATPase
P subunit is not required for catalysis but is necessary for the
structural stabilization of the FI complex. This does not exclude a
role in substrate binding as has been proposed for the homologous
residue of the rat liver F1 0 subunit (Garboczi et al., 1988). In order
to characterize further the role of H211 of the yeast F1 0 subunit,
kinetic analyses of the mutant ATPases were undertaken. First, Km
values for ATP hydrolysis by the mutant enzymes were determined.
This was done to evaluate the contribution of H211 to substrate
binding. Second, the pH optima for ATP hydrolysis, and the pKa
values of the groups involved in catalysis were determined in order
to assess the influence of position 211 on the catalytic mechanism.
Finally, the substrate preferences of the wildtype and mutant
enzymes were qualitatively determined in order to gain insight into
structures of the active sites of the mutant enzymes. In an
additional experiment, the contribution of the 2' hydroxyl group of
ATP to the wildtype catalytic mechanism was investigated by
comparing the kinetic constants of ATP and dATP hydrolysis.

Determination of Kinetic Constants and pKa Values
Km and Vmax for ATP hydrolysis were determined by analysis
of v vs [S] plots. Initial velocities were measured at ten different
substrate concentrations. Each velocity was the average of 3-6
measurements. The Enz-Fitter program was used to fit a curve to the
data and to calculate Km and Vmax using the Michaelis-Menten
The buffer systems used for determination of pH profiles were
50 mM each MES, BES, and BICINE for pH 5.5-9.0, and 50 mM each
PIPES, TRICINE, and CHES for pH 6.25-10.0. At pH values of 5.5 and
6.0 the characteristic lag of the pyruvate kinase-lactate
dehydrogenase coupling system became pronounced. To alleviate
this problem, 8 units of each enzyme were added to each ml of
reaction cocktail immediately before each assay. This ensured the
measurement of initial rates at these pH values. pKa values for
ionizing groups involved in catalysis were calculated using the BELL
program (Cleland, 1979). Stability of ATPases at extremes of pH was
confirmed by incubation of submitochondrial particles at pH 5.5 or
9.0 for 30 minutes at 40 C and subsequent assay of activity at 12 mM
ATP and pH 8.0. The stability of the submitochondrial particle
ATPases used in the determination of pKa was further tested by
incubation in the assay mixture at either pH 5.5 or pH 8.5 and 300 C
for the standard assay length (5 minutes), pH was then adjusted to
6.5 and activity was compared to that of particles which had been

incubated at pH 6.5 for 5 minutes at 300 C. No loss of activity as a
result of incubation at pH 5.5 or 8.5 was detected by this assay.
Measurement of Km Values for ATP Hydrolysis by the Mutant
Submitochondrial particles were prepared from each of the
mutant strains and Km values for ATP hydrolysis were determined
(Table 3.1). Km values for mutant strains PH211D, pH211I, 3H211A,
and pH211K are about 5-fold greater than the wild type. The Km
value for PH211N is about 3.5-fold greater than wild type. This
indicates that H211 of the P subunit contributes to substrate binding
either directly or indirectly. The Km value for pL203F was similar to
wild type. This suggests that position 203 is more likely to be
involved in maintenance of the tertiary structure of the P subunit
than in a binding interaction with the substrate. Km values for the
doubly mutated enzymes were in a range from 0.2 to 0.6 mM. In
addition to the mutations at positions 211 and 203, a substitution of
serine for cysteine at position 32 was also constructed. The aerobic
respiration phenotype and Km for ATP hydrolysis of PC32S were the

same as wild type and this mutation was not studied further.
The mutant ATPases have been shown to be unstable when
removed from Fo, have diminished sensitivity to oligomycin,
assemble to a lesser extent than does the wild type enzyme, and
have elevated Km values relative to the wild type. These results
raise the question of whether the mutant enzymes utilize the same

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EB94RYFLE_NQ245V INGEST_TIME 2012-09-24T14:24:31Z PACKAGE AA00011841_00001