|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
CONTRIBUTIONS OF THE INDIVIDUAL b SUBUNITS TO THE FUNCTION OF
THE PERIPHERAL STALK OF F1Fo ATP SYNTHASE
TAMMY WENG BOHANNON GRABAR
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Tammy Weng Bohannon Grabar
This document is dedicated to my husband, Chuck, and my daughter, Kaia.
The work illustrated in this dissertation and my growth as a scientist could not have
been accomplished without the guidance, encouragement and support of several people
on both the professional and personal levels. The first person I would like to thank is my
mentor, Dr. Brian Cain. He allowed me to join his laboratory when I was fresh out of
college, even though I had no real experiences in a scientific lab. He exhibited extreme
patience while teaching me everything from how to hold and operate a pipette to pouring
agarose gels to cloning my own plasmids. His evident passion and excitement about
science opened my eyes to a whole new world of opportunity. Prior to joining his
laboratory, I had never even dreamed of joining graduate school and pursuing a PhD;
therefore, I feel extreme gratitude and consider myself very fortunate to have j oined his
lab. Once I joined the lab, Dr. Cain allowed me the freedom to make my own initial
scientific and experimental decisions, which was an excellent teaching method for me,
but he was always there for guidance and support whenever it was needed. I would also
like to thank him for being so involved in his lab. On any given day, I knew I could have
his undivided attention if I needed to consult with him. Over the years Dr. Cain has spent
a tremendous amount of time teaching me to think critically about scientific experiments,
how to communicate my data and ideas to others, how to give a professional scientific
presentation, and how to write scientific papers, and for those countless hours I thank
I would also like to thank every person on my committee: Dr. Linda Bloom, Dr.
Art Edison and Dr. Dan Purich from my department, and Dr. Julie Maupin-Furlow from
the Microbiology and Cell Sciences Department. I have known Dr. Maupin-Furlow the
longest. She taught one of the most challenging courses I took as an undergraduate.
When an unexpected death occurred in my family, she was kind enough to allow me to
postpone an exam without questioning my motive, which was very unusual for most of
my professors while I was an undergraduate. Thanks to her, that was the only class in
which my grade was not affected that semester. I would also like to thank her for her
continual support in helping me get into graduate school and then subsequently taking the
time to hike across campus to join my committee meetings. I would like to thank Dr.
Purich for teaching me, in the middle of a physical biochemistry class, that sometimes we
have to take some time off to go outside and get some fresh air. That was always an
important lesson when endless hours in the lab led to careless mistakes. I also enjoyed
his sometimes unusual stories and adventures that he had to share with me when I was
spending entire days in the biochemistry library studying for exams. I would like to
thank Dr. Edison for his constant support and encouragement. He has always been the
first person who publicly and very kindly commended me after my journal club
presentation. I believe his kind words of support through the years helped me to gain the
courage I needed to believe in myself to really deliver a good presentation. I would also
like to thank him for his eagerness to understand every aspect about my proj ect. And
last, but not least, I would like to thank Dr. Bloom. As a woman in my department and a
new mother with a career in academia, she has been a wonderful role model. She has
always had kind words and smiles to bestow on me. I would also like to thank her for
assisting me during my committee meetings when discussions of fluorescence started to
go over my head.
I would also like to take the time to thank everyone that I had the pleasure of
working with in the lab. These are the people I spent countless hours with during the
course of the day and held many scientific and personal conversations with, and I am
happy to be able to call them my friends. I could not have spent the last five years with a
better group of people. Drs. Tammy Otto and Debra Zies were members of the lab when
I first j oined and were the ones who taught me the ways of the lab. Dr. Michelle Gumz
joined the graduate program and subsequently joined Dr. Cain's lab the same time as I. I
would like to thank Tammy, Debbie and Michelle for their scientific and personal support
as well as sharing with me memories of pool barbeques, wedding showers and baby
showers. Dr. Deepa Bhatt joined the lab as a postdoc during my graduate career. Her
friendship and scientific guidance have been very valuable to me. I would like to thank
her for giving me fantastic advice on all of my oral presentations and reviewing all of my
And finally, I would like to thank my family for all the encouragement, love and
support they have unwaveringly offered over the years. I would like to thank my dad for
always finding the positive in everything that was negative and always encouraging me to
overcome the many obstacles that graduate school hurled towards me. He never lost faith
in me, even when I was ready to give up. I would like to thank my mom for her
tremendous support as well. She spent a lot of time and energy stocking my refrigerator
and freezer full of meals when I found that I did not have the time to care for myself. She
has spent many weeks and weekends at my home since my daughter was born so that I
could have extra time to work on my dissertation. And last, but not least, I would like to
thank my husband for his constant emotional support and belief in me. He has been with
me through the thick and thin of graduate school and never once complained of my
emotional torment when things were not going my way. I could not have accomplished
this without him.
TABLE OF CONTENTS
ACKNOWLEDGMENT S .............. .................... iv
LI ST OF T ABLE S ........._..... .......... ............... xii..
LIST OF FIGURES ........._..... ..............xiii..._._. ......
ABBREVIATIONS .............. .................... xvi
AB S TRAC T ......_ ................. ............_........x
1 BACKGROUND AND SIGNIFICANCE ................. ...............1.........._....
Introducti on ............... .... ... ....._..._ .. ... ......._ ..........1
Structure and Function of F1Fo ATP Synthase .............. ...............3.....
The Catalytic Core ................. ...............6........._. ....
The of hexamer ........._... ......___ ...............8....
The y subunit ................. ...............12................
The Rotor Stalk ................. ...............12........... ....
The y subunit ................. ...............14................
The a subunit ........._.. ........... ...............16....
The ring of c subunits ................. ....___ ....___ ...........2
T he Stator Stalk ................. ...............25..............
The a subunit ........._.. ........... ...............27....
The 6 subunit ........._.. ........... ...............33....
The b subunit ........._..._.._ ...._._. ...............37....
Subunit Equivalence ........._..._.._ ...............51.._.._._ .....
F1Fo ATP Synthase Mechanism............... ...............5
Proton Translocation: Driving Rotation .............. ...............57....
Coupling ................. ......... ......... .........6
Catalysis: The Binding Change Mechanism .............. ...............62....
Genetic Expression and Assembly .............. ...............63....
Summary ........._... ...... ._ ._ ...............65....
2 INTEGRATION OF UNEQUAL LENGTH b SUBUNITS INTO F1Fo ATP
SYNTHASE .............. ...............67....
Introducti on ................. ...............67.................
Materials and Methods .............. ...............69....
M material s ................. ...............69.......... .....
Strains and Media ................. ...............70........... ....
Recombinant DNA Techniques............... ...............7
Mutagenesis and Strain Construction ................. ...............75................
Crude Preparative Procedures .............. ...............78....
Determination of Protein Concentration ................. ..............................80
Ni-Resin Purification..................... .... ........8
Assays of F1Fo ATP Synthase Activity .............. ...............82....
Immunoblot Analysis .............. ...............85....
R e sults............... ... .. .. .. ......_ .......... .............8
HA-Epitope Tagged b Subunits................. ................8
Construction and growth characteristics of mutants .............. ...................88
Effects of epitope tags ................ .. .......... .. ...............90.....
Expression of different b subunits in the same cell ........._._............_._.....91
Ni-Resin Purification............... ..............9
V5-Epitope Tagged b Subunits .............. ..........................9
Construction and growth characteristics of mutants .............. ...................97
Effects of epitope tags ................. ...............100...............
Detections of heterodimers. .................... .......... ......_ .............0
Formation of mixed length b subunits in F1Fo ATP synthase ..........................107
Discussion ............ .... __ ...............111..
3 GENETIC COMPLEMENT ATION BETWEEN MUTANT b SUBUNITS IN F1Fo
ATP SY NTHASE ................. ...............114......... ......
Introducti on ................. ...............114................
Materials and Methods ................. ...............116...............
M materials ................ ...............116................
Strains and Media ................. ...............117................
Recombinant DNA Techniques ................. ...............117................
Mutagenesis and Strain Construction ................. ................. ......... ...11
Preparative Procedures ................ ...............120................
Immunoblot Analysis .................... .... ..............12
Assays of F1Fo ATP Synthase Activity .............. ...............121....
Re sults................. .. .... ......... ..... ........ .. ..................2
Construction and Growth Characteristics of Mutants ................. ................. .121
Heterodimer Formation of barg36 Defective Subunits with bwt .......................... 123
Heterodimer formation of bAl534end-his Complemented with bwt-vs ................... 128
Heterodimer Formation of b+124-130-his Complemented with bwt-vs .................... 131
Mutual Complementation ...._._.... .........__........._ ............13
Discussion .........._...._ ......... ...............139.....
4 DEVELOPMENT OF CYSTEINE CHEMICAL MODIFICATIONS OF ALTERED
b SUB UNIT S ................. ................. 143........ .....
Introducti on ................. ...............143................
Materials and Methods .............. ...............146....
M materials ................ ...............146................
Strains and M edia ................. ...............146................
Recombinant DNA Techniques............... ..............14
Mutagenesis and Strain Construction ................. ............. ......... .......148
Crude Preparative Procedures .............. ...............152....
Assays of F1Fo ATP Synthase Activity .............. ...............152....
Immunoblot Analysis .............. ...............153....
Re sults................. .. .... ......... ..... ........ .. ..................5
Construction and Growth Characteristics of Mutants ................. ................. .153
Effects of Cysteine Mutations .............. .....................157
Discussion ................. ...............158................
5 MUTAGENISIS OF THE AMINO AND CARBOXYL TERMINI OF THE b
SUBUNIT INT F1Fo ATP SYNTHASE .............. ...............161....
Introducti on ................. ...............161................
M materials and M ethods .............. ...............165....
M materials ................ ...............165................
Strains and M edia ................. ...............165................
Recombinant DNA Techniques............... ..............16
Mutagenesis and Strain Construction ................. ............. ......... .......16
Crude Preparative Procedures .............. ...............168....
Assays of F1Fo ATP Synthase Activity .............. ...............169....
R e sults............... .. .............. ...............169......
Amino Terminal Mutations ............... .. .. .... ..... .. ............... 169....
Construction and growth characteristics of mutants .............. ................. 169
Effects of amino terminal mutations .............. .....................171
Carboxyl Terminal Mutations ............... ........ ..............17
Construction and growth characteristics of mutants .............. ................. 173
Effects of carboxyl terminal mutation............... ................175
Discussion ................. ...............176................
6 CONCLUSIONS AND FUTURE DIRECTIONS .............. .....................8
C onclusions............... .. .. ..... .. .. ... ...... .................18
Integration of Unequal Length b Subunits into F1Fo ATP Synthase ................. 181
Genetic Complementation between Mutant b Subunits in F1Fo ATP
synthase .............. ... ..... ._ _. .........._.... ... ..... .. .. .. ........ 8
Development of Cysteine Chemical Modifications of Altered b Subunits.......185
Mutagenesis of the Amino and Carboxyl Termini of the b subunit in F1Fo
ATP Synthase ................. ...............186................
Future Directions ................. .. ...... ....... ...............188......
Complementing Mutant b Subunits .................. .... .. ......... .. ...... ._._............8
Function of F1Fo ATP Synthase Incorporated with b Subunit Heterodimers....190
Positions of the Individual b Subunits in F1Fo ATP Synthase...........................190
Length of the Peripheral Stalk in F1Fo ATP Synthase Complexes
Incorporated with Shortened and Lengthened b Subunits ................... ..........191
Other Implications .............. ...............195....
A MUTAGENIC OLIGONULCEOTIDES .............. ...............202....
B DEVELOPING PROTOCOL FOR PURIFYING F1Fo ATP SYNTHASE .............206
Purification of Enzyme Complexes Incorporated with b Subunit Heterodimers .....206
Culture ............ _...... .. ...............206...
Di sruption of Bacteria .............. ...............207....
Ni-Resin Purification................ .............20
V5-Epitope limmunoprecipitation............... ..........1
Detection of Purified Enzyme Complexes ........._.._ ..... ._ ...............2 10
Assays of F1Fo ATP Synthase Activity .........._.._ ......... .............. ....21 1
LIST OF REFERENCES .........._._ .. ...._.. ...............212.....
BIOGRAPHICAL SKETCH .............. ...............237....
LIST OF TABLES
1-1. F1Fo ATP synthase subunit equivalency............... ..............5
2-1. Aerobic growth properties and membrane-associated ATP hydrolysis activity of
mutants expressing epitope tagged uncF(b) genes............... ...............90.
3-1. Aerobic growth properties and membrane-associated ATP hydrolysis activity
of mutants expressing epitope tagged uncF(b) genes ................. .........___.......122
4-1. Description of uncF(b) cysteine mutations .............. ...............149....
4-2. Description of the unc operon cysteine mutations ............. ....................15
4-3. Aerobic growth properties and membrane-associated ATP hydrolysis activity
of mutants expressing cysteine. ................ .............. ......... ........ .....156
5-1. Aerobic growth properties and membrane-associated ATP hydrolysis activity
of mutants expressing uncF(b) mutations at the amino terminus. ................... ....... 170
5-2. Aerobic growth properties and membrane-associated ATP hydrolysis activity
of mutants expressing uncF(b) insertions or deletions throughout the b subunit ..175
6-1. Preliminary data of coexpressed mutant b subunits. ............. .....................190
A-1. Oligonucleotide sequences. ............. ...............203....
A-2. Oligonucleotide description. .............. ...............204....
LIST OF FIGURES
1-1. Timeline of developing views of F1Fo ATP synthase............... ...............5
1-2. Space-filling structural model of Escherichia coli F1Fo ATP synthase............._.._. .....7
1-3. Structure of the a subunit of E coli F1Fo ATP synthase............... ...............19
1-4. Controversial models of the a subunit topology ....._.._.. .... ... .__. ........_.......29
1-5. Amino acid sequence of the E. coli F1Fo ATP synthase b subunit............_..._... ..........3 8
1-6. Gross structure of the E. coli F1Fo ATP synthase and the domains of the b
subunit ........._..._... ...............40.._.._.. .....
1-7. Model for F1Fo ATP synthase peripheral stalk orientation dependent upon the
direction of rotation during ATP synthesis or hydrolysis ..........._.._ ........._..__....47
1-8. Speculative models for the b-like subunits ........._..._... ........._..._...55......_. .
1-9. Model of proton translocation and torque generation in Fo........ .............. .59
1-10. The binding change mechanism .............. ...............63....
2-1. Oligonucleotides for epitope tags and mutagenesis of uncF(b) ................ ...............74
2-2. Construction of the single transcript expression system .............. ....................7
2-3. Histidine and HA-epitope-tagged b subunit expression system ............... ..............89
2-4. Western blot analysis s of hi stidine and HA-epitope tagged b subunits .....................92
2-5. Investigation of detergent solubilization of F1Fo ATP synthase complexes .............94
2-6. Ni-resin purification of F 1Fo, expressing different length b subunits, treated with
the cross-linker BS3 ........._.__ ..... ._._ ...............95...
2-7. Ni-resin purification of histidine and HA-epitope tagged F1Fo treated with the
cross-linker BS3 ................ ...............96........... ....
2-8. Histidine and V5-epitope-tagged b subunit expression system ........._...... ..............98
2-9. ATP-driven energization of membrane vesicles prepared from uncF(b) gene
m utants. ............. ...............101....
2-10. NADH-driven acidification of membrane vesicles prepared from uncF(b)
m utants ................. ...............102.............
2-11. Ni-resin purification ofF FoF ATP synthase treated with the cross-linker BS3.....106
2-12. Ni-resin purification ofF Fo ATP synthase expressing unequal length b
subunits ........._... ...... ..... ...............108....
2-13. Quantitation of b subunit heterodimeric F1Fo ................. ............................110
2-14. Interactions of b subunits of unequal lengths ...._.__... .... .._._.. ........_........112
3-1. Oligonucleotides for epitope tags and C-terminal truncation of uncF(b) ................1 19
3-2. Ni-resin purification ofF FoF ATP synthase incorporated with barg36 Subunit
mutations ....._. ................ ................. 124....
3-3. ATP-driven energization of membrane vesicles prepared from uncF(b) arg36
gene mutants ........... ..... .._ ...............127..
3-4. Ni-resin purification ofF FoF ATP synthase containing a b subunit carboxyl-
terminal truncation .............. ............... 129...
3-5. ATP-driven energization of membrane vesicles incorporated with F,Fo ATP
synthase containing a b subunit carboxyl-terminal truncation................ .............13
3-6. Ni-resin purification of membranes incorporated with b+124-130-his subunit
mutati on ................. ...............132................
3-7. ATP-driven energization of membrane vesicles incorporated with a defective
b+124-130 Subunit mutation ................. ...............133................
3-8. Ni-resin purification ofF FoF ATP synthase incorporated with complementing
defective b subunits ........._._. ._......_.. ...............136....
3-9. ATP-driven energization of membrane vesicles incorporated with F Fo ATP
synthase containing complementing defective b subunits .............. ...................137
3-10. Interactions of defective b subunit with wild type b subunits found in intact
F1Fo ATP synthase complexes .............. ...............140....
3-1 1. Model of F1Fo ATP synthase incorporated with complementing defective b
subunits .............. ...............142....
4-1i. Model of E. coli F 1Fo ATP synthase ................. ...............144...........
4-2. Oligonucleotides for cysteine mutagenesis of the unc operon .............. .... ........._...150
4-3. Expression plasmid of cysteine mutants ................. ...............155........... ..
4-4. Western blot analysis of cysteine mutant b subunits of differing length................. 158
4-5. Model of F1Fo ATP synthase with cysteine substitutions in the b and 6 subunits ..159
5-1 Amino acid sequence and domains of the E. coli b subunit ................. ........_._.....162
5-2. Oligonucleotides for mutagenesis at the amino and carboxyl termini in the unc
operon ................ ...............167.............
5-3. ATP-driven energization of membrane vesicles prepared from b subunit
membrane domain mutants .............. ...............172....
5-4. Amino acid insertion and deletion analysis of the E. coli b subunit ................... ..... 174
5-5. Mutations constructed throughout the b subunit ........._._. ......___ ................177
6-1. Design of FRET experiments to measure the peripheral stalk ..........._..._ ...............193
6-2. Model of rotation inhibition due to a fusion protein on the a subunit ................... .. 195
6-3. Sequence alignments of subunits b and b from various species with the b
subunit of E coli .............. ...............200....
B-1. Diagram of purification procedures for homogeneous heterodimeric bvs/bhis
F1Fo ATP synthase complexes .............. ...............208....
ACMA, 9 -ami no-6 -chl oro-2 -methoxy acri di ne
ADP, adenosine-5' -diphosphate
AO, tegamineoxide WS-35
Apr, ampicillin resistant
ADP, adenosine-5' -diphosphate
ATP, adenosine-5' -triphosphate
b+7-his, SOVen amino acid insertion in the b subunit with a 6X histidine epitope tag at the
ba7-vs, seven amino acid deletion in the b subunit with a V5 epitope tag at the carboxyl
bser84 cys, Substitution of a cysteine for serine at amino acid position 84 in the b subunit
bp, base pair
BS3, bis(3 -sulfo-N-hydroxysuccinimide ester)
BCA, bicinchoninic acid
BSA, bovine serum albumin
Cmr, chloramphenicol resistance
Su2z, the second co-helix in the epsilon (s) subunit
ECD, 1-ethyl-3 [3 -dimethylamino]propyl carbodiimide
ECL, enhanced electrochemiluminescence
EDTA, ethylenediaminetetraacetic acid
FPLC, fast polynucleotide liquid chromatography
FRET, fluorescence resonance energy transfer
g, gravitational force
GFP, green fluorescent protein
HA, peptide epitope of hemagglutinin protein of human influenza virus
ICBR, Interdisciplinary Center for Biotechnology Research
IPTG, isopropyl-1 -thio-P-D-galactopyranoside
LB, Luria Bertani medium
LBG, Luria Bertani media supplemented with 0.2% glucose
LDAO, lauryldimethylamine oxide
LSB, Laemmli sample buffer
MOPS, 3- [N-morpholino]propanesulfonic acid
NADH, P-nicotineamide adenine dinucleotide, reduced form
NFDM, nonfat dry milk
Ni-CAM, high capacity nickel chelate affinity matrix
NMR, nuclear magnetic resonance spectroscopy
PAGE, polyacrylamide gel electrophoresis
PBS, phosphate-buffered saline
PBST, phosphate-buffered saline supplemented with 0.1% tween20
PCR, polymerase chain reaction
Pi, inorganic phosphate
P/O, number of ATP's made per 2 e- transferred to oxygen
PVDF, polyvinylidene fluoride
rms, root mean square
SDS, sodium dodecyl sulfate lauryll sulfate)
TID, 3 -(trifluoromethyl)-3 -(m- [125I]i Odophenyl)di azirine
TBS, tris-buffered saline
TTBS, tris-buffered saline supplemented with 0. 1% tween20
TE, tris[hydoxymethyl]aminomethane, ethylenediaminetetraacetic acid buffer, pH 8.0
Thl, tri s[hy doxym ethyl iami nom ethane, magne sium sulfate buffer, pH 7.5
Tm, melting temperature of double stranded DNA
V5, epitope found in the P and V proteins of the paramyxovirus, SV5
wt, wild type
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
CONTRIBUTIONS OF THE INDIVIDUAL b SUBUNITS TO THE FUNCTION OF
THE PERIPHERAL STALK OF F1Fo ATP SYNTHASE
Tammy Weng Bohannon Grabar
Chair: Brian D. Cain
Major Department: Biochemistry and Molecular Biology
The universal molecule of biological energetic is adenosine triphosphate (ATP),
and the enzyme involved in providing the maj ority of cellular ATP is F1Fo ATP synthase.
Enzymes in this family utilize the electrochemical gradient of protons across membranes
to synthesize ATP from ADP and inorganic phosphate in a coupled reaction. The
cytoplasmic Fl and the membrane-bound Fo sectors are linked by two stalk structures, the
rotor stalk and the peripheral stalk. Proton conduction through the Fo sector drives the
rotation of the rotor stalk within the catalytic core, which is held steadfast by the
peripheral stalk. In Escherichia coli, the 6 subunit of F1 and a parallel homodimer of
identical b subunits constitute the peripheral stalk of F1Fo ATP synthase. Work
accomplished in this dissertation indicates that the bacterial enzyme does not require two
identical b subunits to form the dimer. Two different length b subunits, with a size
difference of at least 14 amino acids, were capable of forming the b dimer of an intact
F1Fo ATP synthase complex. Also, in work presented in this dissertation, a defective
mutation in one region of the b subunit was overcome by dimer formation with a second
b subunit that contained defective mutation in a different region but had a wild-type
sequence in the region of the former defective b subunit. This mutual complementation
between fully defective b subunits indicated that each of the two b subunits makes a
unique contribution to the functions of the peripheral stalk, such that one mutant b
subunit is making up for what the other is lacking. Interestingly, the equivalent of the
bacterial b subunit in plants exists as two genetically different subunits, and the mammal
counterpart exists as at least four subunits. This work suggests that the individual
functions of the b subunits may be reflected in the fact that higher organisms evolved to
encode multiple b-type subunits.
BACKGROUND AND SIGNIFICANCE
The premiere of Peter Mitchell's chemiosmotic theory in 1961 eventually resulted
in the major breakthrough of the characterization of F1Fo ATP synthases. Basically, his
theory stated that protons are pumped across energy transducing membranes, thereby
creating an electrochemical gradient of protons (1). This proton gradient, also known as
the proton-motive force, consists of two components: i) a chemical component due to the
concentration gradient of protons and ii) an electrical component, or membrane potential,
due to the positive charge of the protons (H ). As a result, one side of the membrane is
more positive than the other. The potential energy of this gradient can then be transduced
to chemical energy or utilized to perform work when the protons diffuse back across the
membrane from the higher to the lower potential (2). The protons can diffuse across the
membrane through specific transmembrane proton conductors, which can synthesize
adenosine 5'-triphosphate (ATP) or co-transport solutes, and in the case of bacteria drive
The ability to consume nutrients and convert them to energy is required of all living
organisms, from microscopic bacteria to plants to humans. The universal molecule of
biological energetic is ATP, and in almost all organisms, the central enzyme involved in
providing the maj ority of cellular ATP is F1Fo ATP synthase (3-6). F1Fo ATP synthases
are responsible for the production of ATP in the final step of processes called oxidative
phosphorylation and photophosphorylation. They provide the bulk of cellular energy in
the maj ority of eukaryotes and prokaryotes. The synthesis of ATP occurs at a rate of
about 100 s- which maintains a concentration of about 3 mM ATP in Escherichia coli
and greater concentrations in mitochondria and chloroplasts with no noticeable product
inhibition (3). In eukaryotes, they are located in the inner mitochondrial membrane, or in
the thylakoid membrane of chloroplasts. In most bacteria, F1Fo ATP synthase is located
in the cytoplasmic membrane. Enzymes in this family utilize the electrochemical
gradient of protons across these membranes in order to synthesize ATP from ADP and
inorganic phosphate (Pi) in a coupled reaction. In bacteria, the reaction of ATP synthases
can be reversed if the situation of a dissipated electrochemical proton gradient arises. In
this case, ATP derived from glycolysis can be hydrolyzed in order to pump protons
across the membrane, creating a membrane potential. The membrane potential can then
be utilized to drive other cellular processes such as the extrusion of sodium ions, nutrient
uptake and flagellar rotation.
An explosion of research concerning F1Fo ATP synthase has occurred during the
past few decades. In particular, a great deal of knowledge of the enzyme has been solved
only in the past decade. A plethora of relatively recent reviews concerning every aspect
of F1Fo ATP synthase can be found in the special editions of Journal ofBioenergetics
and Biomembranes (volume 32, 2000) and Biochimica et Biophysica Acta (volume 1458,
2000) as well as reviews authored by Noji and Yoshida (2001), Senior et al. (2002),
Capaldi et al. (2002) and Weber and Senior (2004). This chapter will provide detailed
explanations of what is currently known about F1Fo ATP synthase including the
mechanism of the enzyme as a whole as well as structure and functions of individual
subunits, equivalence of the bacterial enzyme to its eukaryotic equivalents and genetic
expression and assembly. The research presented in this dissertation primarily concerns
the b subunit of the Escherichia coli (E. coli) F1Fo ATP synthase. Hence the b subunit
will be discussed extensively later in this chapter.
Structure and Function of F1Fo ATP Synthase
The structure and function of F1Fo ATP synthases are remarkably similar from
bacteria to humans. In E. coli, the simplest form of the enzyme, F1Fo ATP synthase is a
complex enzyme composed of twenty-two polypeptides of eight different types with the
stoichiometry of u3 38ysab2C10 (Figure 1-2) (6, 7). The deduced molecular size is
approximately 530 kDa. The structure of F1Fo ATP synthase in chloroplasts is very
similar with the exception that there are two isoforms of the b subunit. On the other
hand, the mitochondrial enzyme is more complex, including an extra 7-9 small subunits
which are thought to have roles in enzyme regulation (8-10). Discussion of F1Fo ATP
synthase is commonly divided into two portions, Fl and Fo. The Fl portion of the enzyme
is composed of the cytoplasmic subunits, a3 3ysS, and is responsible for the synthesis of
ATP. The Fo portion consists of the membrane-bound subunits, ab2C10, and is responsible
for the translocation of protons through the membrane. New insights concerning the
functions and the intersubunit contacts have refined the way F1Fo ATP synthase is
perceived, dividing discussions of the enzyme into the catalytic core, the rotor (central)
stalk and the stator (peripheral) stalk (3, 11-14). The catalytic core consists primarily of
the u3 37 subunits, the rotor stalk consists of the ysclo subunits and the stator stalk
consists of the b26 Subunits.
Over the years, F1Fo ATP synthase has received worldwide recognition as the
tiniest rotary motor known to mankind (15, 16). Protons passing through the enzyme
complex drive the rotation of the rotor ysclo subunits at about 100 Hz. This rotation,
which is absolutely essential for the machinery of the enzyme, transmits energy over a
distance greater than 100 A+ by providing the means by which conformational changes in
the Fl catalytic core, u3 3, take place for the synthesis of ATP (3, 4).
Structural studies of F1Fo ATP synthase commenced in the early 1960's and persist
to this day in pursuit of a complete high-resolution structure. Negative staining
procedures in the early 1960's initially revealed the traditional tripartite features of the
enzyme complex from sub-mitochondrial particles, consisting of what was referred to as
the headpiece, stalk and basepiece (Figure 1-1) (17). Ten years later, the first electron
micrograph of a detergent-solubilized F1Fo ATP synthase was published, confirming the
existing idea of a tripartite molecule (18). Appreciably, electron microscopy (EM) in
combination with other biochemical data of isolated Fl exposed a hexagonal arrangement
of alternating subunits with a seventh mass found in the center of the array (Figure 1-1)
(7, 19). Based on this premise it was first suggested that Fl consisted of an alternating
hexagonal array of three a and three P subunits with the y, 6 and a situated centrally (7).
The idea did not gain favorable recognition for some twenty years until verified by x-ray
crystallography (20, 21). Continuous improvements in EM technology led to numerous
publications of various ATP synthases which defined the average overall dimensions of
about 190 A+ from top to bottom and about 37 A+ assigned to the stalk structure (22-25).
Using a combination of traditional biochemical, molecular biological and immunological
techniques along with EM, many important discoveries were made that led to what we
now understand of F1Fo ATP synthase. The first direct evidence for rotation of the stalk
appeared in 1990 (24), but was not followed by the visualization of the peripheral stalk,
of rat F, at
Figure 1-1. Timeline of developing views of F1Fo ATP synthase. Electron microscopy
and biochemical analysis from the early 1970's through the 1980's allowed
visualization of the classical tripartite features of F1Fo ATP synthase
consisting of what was referred to as the headpiece, stalk and basepiece.
Furthermore, the arrangement of the F1 subunits were first proposed in 1974,
though it did not gain favor until high resolution structure was obtained
twenty years later. The first direct evidence for rotational catalysis appeared
in 1990. Improved EM techniques showed the existence of a peripheral stalk,
assigned the role of the "stator" to hold the F1 sector in place against the
proposed rotation, and a "cap" in the late 1990's. In 1994, the first high-
resolution structure (2.8 A+) of F1 appeared, consisting of the OC3 3 hexamer
with partial structure of the y subunit. Currently there is no high resolution
structure for the entire F1Fo ATP synthase complex though many of the
subunits have been solved individually or in part by model polypeptides.
at 2.8 A
assigned the task of the "stator" to hold Fl in place against the rotation of the centrally
located stalk, until many years later (26-29).
Today there is still no high-resolution structure of the entire F1Fo ATP synthase
enzyme complex from any organism. X-ray crystallographic and NMR data of partial
complex structures from rat, bovine, yeast and E. coli or model polypeptides deduced
from nucleotide sequences have accumulated over the past twenty years to allow for a
composite structural model with both high- and low-resolution structures (Figure 1-2).
Currently, complete high-resolution structures are available for the a, P, y, E and c
subunits and partial structures for the 6 and b subunit. There is currently no high-
resolution data for the membrane-integral a subunit.
The Catalytic Core
The first high-resolution structure, resolved to 2.85 A+, consisted of a3 3Y Of Fl
prepared from bovine heart under inhibited conditions and in the absence of Pi and
substoichiometric amounts of ADP (20). This major breakthrough was shortly followed
by a 2.80 A+ Fl isolated from rat liver in the absence of the physiological cation, Mg2+
(21). The arrangement of the subunits in the two structures obtained were exceptionally
similar and confirmed Catteral and Pedersen's proposal made two decades prior by
showing the three a and three P subunits arranged alternatively with the amino and
carboxyl-termini of the y subunit, each forming an a-helix, extending up through the
center of the hexamer (Figure 1-2) The only difference in the structures, which was the
occupancy of the three nucleotide binding sites located at the up interfaces, was likely
due to the difference in preparation conditions (7) or crystal quality. In the former,
Figure 1-2. Space-filling structural model ofEscherichia coli F1Fo ATP synthase. The
model is based on a composite of high and low resolution structures taken
from E. coli, yeast and bovine F1Fo ATP synthases. F1Fo subunits included in
the model were u3 38ysab2C10. The subunits were color coded as follows: a,
red; p, green; y, cyan; s, orange; 6, yellow; b, blue; a, brighter yellow; c,
darker blue. The direction of the arrow indicates the direction of rotation of
ysclo during ATP synthesis. The yellow and blue cylinders represent the a
subunit and portions of the b subunit that currently have no high-resolution
structures from any species.
one of the catalytic sites was empty and in the latter, all three active sites were occupied
with nucleotides (20, 21). Since the occupancy state of the three catalytic sites has been
of considerable debate, a more accurate depiction of the Fl moiety, which would give
some insight about the mechanism of ATP synthesis, must be attained from crystals
obtained under physiological conditions. A more accurate depiction of the structures and
roles of each individual subunits of the Fl ATPase follows.
The up hexamer
Homology. The a subunit of the E. coli F1Fo ATP synthase, product of the uncA
gene, is the largest subunit consisting of 513 amino acids with a deduced molecular
weight of 55,3 13 Da. The P subunit, a product of the uncD gene, is a 459 amino acid
subunit, with a molecular weight of 50,325 Da. Based on the primary sequences, the a
and p subunits of E coli Fl have the most obvious homologies in the chlororplast and
mitochondrial enzymes (30). The highest conserved subunit from the E. coli F1Fo ATP
synthase is the p subunit with approximately 70% homology with the chloroplast and
mitochondria equivalents (31). The a subunits exhibit roughly 50% homology (31). A
total of 6 nucleotide binding sites are housed at the up interfaces, three catalytic
contributed primarily by the P subunit and three noncatalytic housed primarily by the a
subunit (32, 33). The nucleotide binding regions have sequence homologies with other
proteins that bind nucleotide or phosphate, including secA protein, N-ethylmaleimide
sensitive fusion protein, herpes simplex virus UL15, Ca2+-ATPase, H /K+ ATPase and
Na /K+ ATPase (34-37). Furthermore, the nucleotide binding motif, GXXXXGKT/S,
known as the Walker A motif, which was first identified in the a and P sequences ofF1,
has been found to be conserved in the high-resolution structures of other proteins
including p21'", adenylate kinase, RecA, elongation factor Tu, and transducin-a (20, 38-
Tertiary structure. The first high-resolution structure of bovine Fl resolved at the
atomic level (2.8 A+) was solved by Walker's group a decade ago (20). It was found to be
a flattened sphere approximately 80 A+ high and 100 A+ wide with the three a and three P
subunits arranged as a hexamer of alternating subunits around a centrally located 90 A~
long a-helix formed by the y subunit. A dimple 15 A+ deep is located at the top of F1.
The amino-terminal regions of the a and P subunits were once thought to be in close
proximity to the membrane due to labeling experiments (43). Contrary to this early data,
the crystallographic data placed the amino-terminal regions on the top of the u3 3
hexamer over 100 A+ away from the lipid bilayer.
The folds of the a and P subunits were found to be nearly identical. They each
consisted of a six-stranded P-barrel at the amino terminus (al9-95, P9-82), ai central a-p
domain containing the nucleotide-binding site (u96-379, P83-363) and a bundle of seven and
six helices at the carboxyl termini of the a and P subunits, respectively (u380-510, P364-474)
(20). The nucleotide binding domain consisted of a nine stranded P-sheet with nine
associated a-helices, of which the a-carbons of the seven P-strands and the seven
associated helices can be superimposed onto the RecA protein ATP binding site with an
rms separation of 1.9 A+ (20).
The three catalytic sites were located at the interfaces of the u3 3 hexamer. In the
original crystal structure, now commonly referred to as the reference structure, two of the
three sites were occupied by nucleotide, containing MgADP ("(PDP Site") and MgAMP-
PN~P ("(PTP Site"). The third site was empty and designated "(PE." The PDP and PTP
subunits were in similar, closed conformations whereas the PE adopted an open
conformation, differing from the other two by a large hinge motion of the carboxyl
terminal domain of greater that 20 A+. Subsequently, several high-resolution structures of
crystals obtained under various nucleotide conditions gave the same overall structure of
bovine Fl with two nucleotides bound ("two nucleotide structures") (44-49). The PTP Site
was found to occasionally contain a diphosphate nucleotide, establishing that there is no
requirement for PTP to be occupied by a nucleotide triphoshate to produce a
conformational change (48). A more recent structure of bovine Fl solved by the Walker
group at 2.0 A+ showed all three catalytic sites bound by nucleotide (50). Both the PTP
and PDP Sites contained MgADP, adopting the closed conformation, whereas the site
corresponding to the PE Site in previous structures contained MgADP+Pi and adopted a
half-closed conformation. It is thought that the PDP Site is actually the catalytic site. The
structure of rat liver Fl was solved (2.8 A+) in the presence of physiological concentrations
of nucleotides but in the absence of the physiological cation, Mg2+. In this structure, all
three nucleotide binding sites adopted strikingly similar conformations, analogous to the
PDP and PTP Of the previously reported structure of the bovine Fl. This structure had no
indication of the open conformation and showed the presence of nucleotide in all three
sites (21). The structure of an OE3s 3COmplex from a thermophilic bacterium was solved
in the absence of nucleotides and exhibited all three P subunits in the open conformation,
suggesting there is a correlation between the open conformation and the absence of
nucleotide (5 1). A low resolution crystal structure (4.4 A+) of the E. coli Fl has been
obtained by Capaldi's group, in which the catalytic sites are thought to be very similar to
that of the bovine structure; however, the occupancy state of the nucleotide binding sites
was unclear (52). The frequent reports of "two nucleotide" structures have bewildered
scientists due to the vast body of biochemical data from numerous laboratories, using a
variety of techniques, which establish indisputably that all three catalytic sites are readily
filled with nucleotide (32, 53). It is possible that the enzyme preferentially crystallizes in
a ground state intermediate which may occur after the release of product, leaving one site
empty and opened (54). Nevertheless, the accumulating structural data may be indicative
of several intermediary steps that may form during the synthesis of ATP. A detailed
account of the mechanism of ATP synthesis follows later in this chapter.
The crystal structure does offer some insight of the chemical mechanism of ATP
synthesis (20). In the P subunit, 4.4 A+ from the terminal phosphate of the bound
nucleotide triphosphate, there is clearly a density for a water molecule hydrogen bonded
to the carboxylate of PglU1SS. This carboxylate is positioned to allow an inline
nucleophilic attack of the water molecule on the terminal phosphate. The guanidinium of
a neighboring residue, aarg373, iS thought to help stabilize the negative charge on the
terminal phosphate during the transition state (20). This same arrangement can be found
in the catalytic site of transducin-a (42). The crystal structure also provides some insight
as to why the nucleotide binding sites in the a subunit are noncatalytic. There is no
spacial equivalent of the carboxylate of PglU1SS in the a subunit. The spatial equivalence
in the a subunit is filled by a agln208, with the side chain pointed away from the terminal
phosphate (20). The binding of the adenine to the noncatalytic site of the a subunit is
highly specific, unlike the P subunit nucleotide binding site, which can accommodate
GTP, ITP as well as ATP (55, 56). This specificity is due to several hydrogen bonds as
well as the presence of the Ptyr368 ClOse to the 2-position of the adenine ring in the a
subunit, while in the P subunit the adenine is in contact with a hydrophobic interface
(20). Though the binding sites in the a subunits are highly specific, the roles remain
The y subunit
The y subunit plays an important role in the catalytic core. Interactions between the
amino- and carboxyl-terminal a-helices and the u3~ 3Subunits are responsible for the
conformational changes that result in ATP catalysis. The y subunit is a fundamental part
of the rotor stalk.
The Rotor Stalk
Two narrow stalks, a centrally located stalk and a peripheral stalk, have been
observed to link the catalytic core of F1 and the membrane-bound proton translocating Fo
with about of 40-45 A+ in between (27). The central stalk came into view three decades
ago via EM and has since become widely referred to as the rotor stalk. The rotor stalk
consists of the y and a subunits. The bottom of the rotor stalk is connected firmly to the
Fo ring of c subunits located in the membrane and the top extends 90 A+ within the u3 3
hexamer of F1 where it forms crucial interactions with both the a and P subunits (57-59).
F1Fo ATP synthase is an extraordinary enzyme due to its ability to couple potential
energy, obtained from proton translocation through Fo in the membrane, to the synthesis
of chemical energy, over 100 A+ away in Fl, by a rapid rotation of subunits. Although
predicted by Boyer in the 1970's, evidence of rotation did not appear until the early
1990's. The X-ray crystal structure solved by Walker's group suggested that the y
subunit was the rotating subunit by suggesting it could distribute itself to all three P
subunits as opposed to just one (20). Consistent with this idea was inhibition of the Fl
complex by crosslinking the y subunit to one of the a or P subunits (60, 61) and recovery
after photobleaching experiments (62, 63). More convincing evidence was provided
when Duncan et al. crosslinked the y subunit to an unlabeled P subunit by disulfide bond
and then mixed the y-P complex with 35S-labeled P subunit (along with the a, 6 and a
subunits). When the disulfide bond was broken and ATP was added, the y subunit was
observed to switch from labeled to unlabeled P subunit (64). Finally, direct evidence was
achieved in single molecule experiments by attachment of a fluorescent actin filament to
the y subunit and observance of unidirectional rotation of the actin filament upon addition
of ATP (15). Direct observation of the rotating y subunit was soon followed by
observance of the rotation of the s and c subunits at the same speed and direction,
indicating that these three subunits rotate in synchrony, forming the central rotary
machinery of the enzyme complex (65-67). Until very recently, rotation has only been
observed in the direction of ATP hydrolysis. Direct evidence for the synthesis of ATP by
Fl has been shown by attaching a magnetic bead to the y subunit ofF1 fixed to a glass
surface and the rotating the bead, in the appropriate direction, using electrical magnets
The first structural information obtained for the a subunit ofE coli was
accomplished by nuclear magnetic resonance (NMR) studies (69) and is good agreement
with the crystal structure solved at 2.3 A (70). In all previous crystal structures of the
rotor stalk, the portion of the rotor stalk' s y subunit protruding from the Fl ap hexamer
and the a subunit were disordered. Recently, the structure of the bovine homologs of the
rotor stalk y and a subunits has been solved and refined to 2.4 A+ (48). The structures of
the E. coli y and a subunits are remarkably similar with that from bovine Fl. When
comparing the structures of the y and a rotor stalk obtained under different conditions or
from different sources, in combination with an overwhelming amount of biochemical and
immunological evidence, it is clear that the domains of the two subunits undergo maj or
shifts in position, which reflects its fundamental role in the synthesis of ATP (20, 44, 48,
The y subunit
In E. coli, the y subunit is the third largest subunit of F1Fo ATP synthase, encoded
by the uncG gene as a 286 residue polypeptide with a deduced molecular weight of
3 1,563 Da. It plays an essential role in coupling proton transport to the synthesis of ATP.
The first visualization of a portion of the y subunit was a bovine Fl partial structure
solved in combination along with the u3 3 hexamer revealing three a-helices (20). The
Y209-272 (TOSidues 223-286 in the E. coli sequence) carboxyl terminus formed a long (90 A+)
a-helix extending from the stalk structure seen by EM to about 15 A+ from the top of the
hexamer. The bottom half of this helix formed a left-handed anti-parallel coiled coil with
a shorter a-helix composed of the amino-terminal residues Yl-45 (20). The two helices
protruded about 30 A+ from the bottom of F1. An approximately 200 kink in the latter
helix was produced by ypro40 and a similar but less pronounced kink was induced by Yleu217
in the former (48). A third, much smaller a-helix, composed of y73-90 (TOSidues 83-99 in
the E. coli sequence) was inclined at about 45 degrees from the larger helices and located
directly under the Fl hexamer.
More recently, the complete structure of the bovine rotor stalk has been solved to
2.4 A+ (48). The overall length of the stalk, from the carboxyl terminus of the y subunit to
the very bottom where it contacts the ring of c subunits, was 1 14 A+. The portion that
protrudes from the u3 3 hexamer, i.e. the part seen in electron micrographs, was 47 A~
long and 54 A+ wide at its largest cross-section. A completely new a/0 domain,
consisting of a five-strand P-sheet (1-5) and six a-helices (a-f), was identified in the
complete structure of the y subunit. Helices a and f extended into the u3 3 hexamer to
form the antiparallel coiled coil discussed previously. Strands 1-3 along with helices b
and c formed a Rossman fold that forms extensive interactions with the a subunit as well
as the ring of c subunits in the membrane (discussed below). This fold was linked to a P-
hairpin, formed by strands 4 and 5, by helix d. Overall, this a/0 domain had a globular,
oval shape with the dimensions 5 1 A+ wide by 41 A+ high. The positioning of the a/p
domain at the base of the rotor stalk may provide stability to the structure of the rotor
stalk during rotational catalysis (48).
A low-resolution crystal structure of the E. coli F1Fo ATP synthase y subunit was
solved to 4.4 A (52). Upon comparison with the high-resolution bovine structure
obtained one year later, a few differences were observed (48). Helices a and f (see
above) were extended by and extra 12 and 20 residues, respectively. Four additional
putative a-helices, designated B and D-F, were found in the E. coli structure and had
little agreement with the bovine y subunit structure. E. coli helix B runs parallel with the
bovine P-strand 5 and may correspond to it. Helix D has no apparent equivalent in the
bovine model and helices E and F appear to overlap with regions in the bovine 6
(bacterial s, see Table 1-1) and a (no equivalent in bacteria) subunits. The remainder of
the E. coli structure appeared similar to the bovine structure.
The crystal structure displays a strikingly asymmetrical Fl due to differences in the
domains of the a and p subunits and the interactions formed with the single y subunit
(20). The obvious asymmetric positioning of the coiled coil of the y subunit is a key
feature to the mechanics of the binding change mechanism of F1Fo ATP synthase. Its
large carboxyl terminus a-helix passes through a hydrophobic sleeve formed by six
proline-rich loops of the a and P subunits, undoubtingly resulting in the conformational
changes occurring in the catalytic sites (20). In the PE Subunit (see above), several
hydrogen bonds are formed with the y subunit, which forms a "catch", resulting in
conformational changes. Specifically, residues Yarg254 and ygln255 in the carboxyl terminal
helix form hydrogen bongs with PE-asp317, PE-thr318 and PE-asp319. Also, a second "catch" is
formed between the carboxyl terminal domain of the PT Subunit and the short helix of the
y subunit. Hydrogen bonds formed between Yylss, Yys,90 and Yalnso with PT-asp394 and PT-
glu398. This sequence of the P subunit, DELSEED (P394- 400), iS a portion of the binding
site of amphipathic cationic inhibitors and putatively the ATPase inhibitor protein (81-
83). Recently, mutations of residues involved in the "catch" loops were shown in inhibit
ATP hydrolysis activity by the soluble F1-ATPase (84). Structural information suggests
the two antiparallel coiled coil a-helices of the y subunit may unwind during rotational
catalysis and the a subunit rotates around the Fl axis while undertaking a net translation
of about 23 A+ (85). It is likely that these gross changes observed in the structures
revealed individual functional states of the enzyme complex during catalysis.
The E subunit
The a subunit ofE. coli F1Fo ATP synthase consists of 138 amino acids with a
molecular weight of 15, 068 Da and is encoded by the uncC gene. The a subunit has
several putative functions in the F1Fo ATP synthase complex including structural,
inhibitory and coupling roles. Structurally, the binding of F1 to Fo has long been known
to require the presence of the a subunit, which had implicated it as part of stalk structure
(86). In isolated Fl, and in isolated F1Fo to a lesser extent, it has been shown to inhibit
ATP hydrolysis activity (87-89). The removal of a from isolated Fl resulted in up to a
10-fold increase in ATP hydrolysis activity. Furthermore, a truncated version of the E
subunit lost all inhibitory functions but still promoted binding of F1 to Fo, hence the
inhibitory feature has been assigned to the extreme carboxyl terminus (90). It was
speculated that it acts as an inhibitor by reducing the rate at which the product is released
from the catalytic site (91). Also, the a subunit, as part of the rotor stalk, plays a role in
the coupling of proton translocation to the catalytic site. The a subunit' s diversity of
functions is supported by the findings that it produces several points of interactions with
the oc (61, 74, 92), P (61, 88, 89, 93, 94) and y (74, 95-97) subunits of F1 and the c (98,
99) subunits of Fo.
An innovative set of experiments conducted by members of the Dunn laboratory
made use of different sized fluorescent proteins, ranging from the 12 kDa cytochrome
b562 prOtein to the 30 kDa flavodoxin reductase protein with a 20 residue linker (100).
The proteins were fused to the carboxyl-terminus of the a subunit. Since the a subunit is
part of the rotor stalk, according to the concept of rotational catalysis, the fusion of a
large protein at this site should sterically hinder rotation due to the presence of the
peripheral stalk. Cells expressing the smaller cytochrome b562 prOtein fused to the E
subunit grew on minimal media, indicating a functional F1Fo ATP synthase complex.
However, cells expressing the larger flavodoxin reductase protein fused to the a subunit,
though found in an intact enzyme, failed to grow. These results provided the first
evidence, in vivo, supporting rotational catalysis.
High-resolution structural data of the E. cobi s subunit was first solved by NMR
followed by the X-ray crystal structures of the isolated E and completed ye subunits (69,
70, 101). The isolated a subunit consisted of an 84 residue amino-terminal P-sandwich
domain and a 48 residue carboxyl-terminal helix-turn-helix domain in which the two a-
helices formed an antiparallel hairpin (70). The P-sandwich consisted of two five-
stranded P-sheets folded as a rigid, flattened P-barrel. The structure of the isolated a
subunit from E. cobi was very similar to the Fl complex isolated from bovine, which
included the u3 378 subunits (48, 70). Superimposition of 127 of the amino acid Ca
resulted in an rms deviation of 1.6 A+. On the other hand, in the E. cobi y-s complex,
resolved to 2. 1 A+, the a subunit assumed a strikingly different conformation, in which the
two a-helices of the antiparallel hairpin at the carboxyl-terminus are wide apart and
wrapped around the y subunit (Figure 1-3A) (101). Subsequently, both conformations of
the a subunit have been trapped in E. cobi F1Fo ATP synthase by crosslinking
experiments, confirming the existence of both in an intact enzyme complex (102).
Furthermore, Capaldi's group observed that when the carboxyl-terminal helices assume
the hairpin conformation, bringing them closest to the Fo sector, ATP hydrolysis was
activated. Still, the enzyme was fully coupled in the direction of either hydrolysis or
synthesis. In contrast, when the two helices were open, assuming a position closer to the
Fl sector, ATP hydrolysis was inhibited and the enzyme functioned only in the direction
Figure 1-3. Structure of the a subunit of E coli F1Fo ATP synthase. The residue numbers
and subunit labels are color coded to match the subunits it represents. The E
subunit has been suggested to undergo large conformational changes during
catalysis from an overwhelming amount of biochemical data. Two different
structures have been obtained for the a subunit, confirming the previous data.
A) Superimposition of the Ca trace of the E structure obtained from isolated a
subunit (u-helices shown in red and the P-sandwich shown in blue) with the
structure obtained from the y-s complex (yellow). B) Composite structural
model. Rotor stalk is based on the crystal structure obtained from the
y-s complex. The PDELSEED, Ecu2 and the s P-sandwich are indicated to show
the close proximity of PDELSEED and Su2 aS well as the relative distance
between the Su2z and the s P-sandwich.
of ATP synthesis. This conformational switch of the a subunit was therefore suggested to
play a key role as a selective inhibitor of ATP hydrolysis and directional regulator of
rotational catalysis by acting as a ratchet (102).
Movement of the two E a-helices was consistent with other observations. Changes
in the a subunit conformation due to nucleotide occupancy in the catalytic sites has been
observed in tryptic proteolysis experiments (89). Cysteine replacements in the carboxyl-
terminal a-helix (Su2) TOSulted in crosslinks with the a and P subunit (61, 103). More
importantly, treatment with a zero-length crosslinker, 1-ethyl-3 [3-dimethylamino]propyl
carbodiimide (EDC), resulted in a high yield of crosslinks between the a subunit and the
DELSEED (P380- P386) TegiOn of the P subunit (PDELSEED) following ATP hydrolysis in
the catalytic sites, but these interactions are disrupted upon the subsequent binding of
ATP. Also, in a composite structure of F1Fo ATP synthase incorporated with the E. coli
e-y complex as solved by Rodgers and Wilce, the s P-sandwich was at least 10 A+ away
from the PDELSEED TegiOn (Figure 1-3B) (101). The carboxyl-terminal Su2 prOduces
several points of interactions with the a, p and y as well as points of interactions with its
own P-sandwich domain a (61, 74, 88, 89, 92-99). In order for the Su2 to interact with
the PDELSEED, Ecul and the a subunit P-sandwich domain, it is clear from the structure that
the a subunit would be required to undergo large movements during the catalytic cycle.
In E. coli, F1Fo ATP synthase can act in two functional directions. In the case of a
dissipated electrochemical gradient, the F1Fo complex acts primarily as an ATPase in
order to pump protons across the membrane to provide a gradient to drive various ion
transport activities in the cell. Under severe conditions where cellular ATP levels are
exceedingly low the enzyme acts predominantly in the direction of ATP synthesis.
Therefore, one can imagine that the ability to selectively turn off ATP hydrolysis, while
preserving ATP synthesis function, may be important for E. coli. In mitochondria the
ability to control the F1Fo ATP synthase complex is essential. It is believed to act
exclusively in the direction of ATP synthesis and is strictly regulated (104).
The ring of c subunits
The c subunit is one of the three membrane-bound Fo subunits of F1Fo ATP
synthase. Ten copies of the c subunit form a ring in the membrane that plays a crucial
role in both proton translocation and rotation of the rotor stalk (105). It is the smallest
subunit of the F1Fo ATP synthase enzyme complex with 79 amino acids and a molecular
weight of 8,256 Da and it is encoded by the uncE gene.
Structure and topology. Early biochemical, genetic and immunological data had
suggested the structure of the c subunit to be that of a helical hairpin with two lipophilic
co-helices (amino acids 1-41 and 50-79) separated by a hydrophilic loop (amino acids 42-
49). Both of the putative transmembrane helices in the regions of Cleu4-leul9 and cphe53-phe76
were vulnerable to chemical modification by the nonpolar photoreactive reagent 3-
(trifluoromethyl)-3 -(m- [125I]i Odophenyl)di azirine (TID), whi ch i s a hydrophobi c carb ene
generator that is believed to react from the nonpolar region of the lipid bilayer, indicating
that these regions were in fact in the hydrophobic phase of the bilayer (106). The loop
region of the hairpin is substantially more polar and antibodies against it were shown to
bind to F1-stripped inverted membrane vesicles suggesting that it resides in the
cytoplasmic of the cell (107, 108). Both of the transmembrane helices are devoid of
charged amino acids with the noteworthy exception of Casp61 in the center of the second
helix, which undergoes a protonation and deprotonation cycle during proton translocation
(discussed in "F1Fo ATP Synthase Mechanism" below) Dicyclohexylcarbodiimide
(DCCD) reacts specifically with casp61, blocking proton translocation, and this reaction is
blocked by the mutationS Cala24-ser Of Cile28-thr, Suggesting that the c subunit is folded in
such a way so that the asp61 of the second helix is in close vicinity to residues 24 and 28
of the first helix (109, 110). This model was further supported by the ability to move the
critical aspartate from residue 61 in the second helix to residue 24 in the first helix
without disruption of enzyme function (1 11). Also, only one subunit in the ring of 10 c
subunits need be modified by DCCD to inhibit activity, indicating that each one of the c
subunits is consecutively involved in proton translocation (112, 113). Furthermore,
modification of the c subunit by DCCD trapped the configuration of the a subunit
(discussed above), providing evidence for a connection between the c and a subunits (89).
Intersubunit contacts made by the c subunit are evident from mutational data and
gives some insight to the topology. Mutations constructed in the polar loop region can
disrupt the binding of F1 to Fo (114-117). Three conserved amino acids, carg41-Cgln42-Cpro43,
lie at the apex of the polar loop region and are predicted to interact with the F1 s subunit
(98, 116). F1Fo ATP synthase complexes with the uncoupling mutation, Cgln42-glu, were
found to be recoupled with a second site suppressor mutation in the a subunit of F1,
Sglu31-gly, val, or cys (98) and was shortly followed by the observance of disulfide bridge
formation between the c subunit and the s and y subunits (99, 118). Also, switching the
essential aspartate from residue 61 in the second helix of the c subunit to residue 24 of
the first helix (discussed above) resulted in a functional F1Fo ATP synthase complex
though the cells were not as healthy compared to cells containing a wild-type enzyme
complex (111). Eighteen third-site suppressor mutants were found that helped to
optimize this cala24-asp,asp61-gly defect, with only five laying on the c subunit and 13 in the
a subunit, all near the aarg210 TOSidue, which is required for proton translocation (further
discussed below) (119, 120). Early models of the organization of Fo suggested that the
ring of c subunits were situated on the periphery, surrounding the centrally located a and
b subunits, which rotated in the center of the ring (121). This model was proved wrong
by high-resolution NMR data (122) and cross-linking experiments (123, 124) which
indicated that the oligomer of c subunits are closely packed with a lipid filled core less
than 25 A+ wide. The individual c subunits are packed front-to-back such that the second
helix of each is situated towards the exterior and the first helix is located on the interior,
which renders the casp61 exposed to the lipid environment. The uncommonly high pKa
(7. 1) of the casp61 carboxyl side chain is likely due to this hydrophobic environment (125).
Furthermore, scanning force and cryoelectron microscopy demonstrated that Fo is
asymmetrically arranged in the membrane (27, 126, 127). For these reasons, the a and b
subunits are thought to be situated to the periphery of the ring of c subunits.
High resolution structures of membrane-bound proteins were nonexistent for many
decades past structural determination of soluble proteins and still prove difficult to this
day due to their highly hydrophobic nature. The membrane intrinsic c subunit of E coli,
which was solved by NMR in an organic solvent (chloroform-methanol-water) in the
1990's, was one of the earliest high-resolution structures of a transmembrane helical
protein (122, 128-130). Notable, the c subunit could be reconstituted from the organic
solvent mixture with complete preservation of function; therefore, it was clearly not
irreversibly denatured (113). As predicted two decades prior, the c subunit folds as a
hairpin of two extended co-helices with the casp61 Of the second helix packed less than 5 A~
from the cala24 and cile28 Of the first helix (122). With the exception of Casp61 in the second
helix, both helices consist entirely of nonpolar amino acids. The first helix is greatly
enriched in glycines and alanines, which led to a smaller diameter. The a-helical
structure of the second helix is interrupted around casp61 due to disrupted hydrogen bonds
around cpro64 which cause the angle of the helical packing to change direction from there
to the carboxyl terminus (122). A recent study, using parallax analysis of fluorescence
quenching, the proton binding site casp61 WAS found to be deeply embedded in the
membrane at about 1.8 A+ from the center of the bilayer (13 1).
Stoichiometry. The stoichiometry of c subunits would be valuable in determining
the number of protons transported per ATP synthesized and will directly relate to the P/O
ratio of oxidative phosphorylation. However, the number of c subunits in Fo had been a
matter of controversy for many years. The number of c subunits in an F1Fo ATP synthase
complex was suggested to be between 9 and 14, but whether this number fluctuated based
on the species or environmental conditions or whether it was a fixed number were the
two prevailing arguments until just a few years prior. Based on a related family of
vacuolar (V-type) ATPases, in which the proposed subunit c had evolved into a fused
dimer of four transmembrane helices with a single proton-transporting glutamate in the
center of the fourth helix, Fillingame et al. set out to genetically fuse the E. coli c subunit
by introducing a flexible loop of similar length (123). The generated c-c dimers and c-c-c
trimers resulted in functional enzyme complexes. In combination with crosslinking
studies and normalization to the a/p content of the membranes, the favored stoichiometry
was fixed to 12 c subunits per F1Fo ATP synthase complex (132). More recently, the
experiment was revised to include only trimers and tetramers of the c subunit (105).
Partial activity was observed in complexes incorporated with eight (c4 C4) Of Hine C (C3
+ c3) Subunits and crosslinked products of more than 10 c subunits were observed but did
not purify in intact enzyme complexes. Crosslinking showed that the preferred
stoichiometry of c subunits in intact E. coli F1Fo ATP synthase was c4 C3 C3, Or 10 c
subunits. This number is consistent with the clo oligomer found in the yeast crystal
structure of a yeast F1Fo ATP synthase consisting of a3 3y8C10 TOSOlVed to 3.9 A+ (58).
However, the preferred number may still vary in different species. The archaebacteria
M\~ethanococcus janna;scjii ATP synthase has a natural c subunit trimer and therefore
cannot incorporate the E. coli equivalent of clo in the membrane (133). With 10 c
subunits present in the membrane 3.3 protons are required per ATP synthesized, which
was compatible with the early experimentally determined ratio of 3 H /ATP estimated
from E. coli whole cells (134). This value also indicates a P/O value of 2.3 from NADH-
linked substrates and 1.4 for succinate, also compatible with the predicted values of 3 and
2, respectively (135).
The Stator Stalk
The b and 6 subunits were once believed to form the central, rotating stalk of F1Fo
ATP synthase. However, high resolution crystallographic data refuted this idea in the
mid 1990's (20). The stator stalk did not come into view until improved EM technology
observed a peripheral stalk in the late 1990's (26-29) and the visualization of the 6
subunit, as a "cap" structure atop F1Fo ATP synthase, soon followed (27-29). As its name
implies, the role of the stator is to hold the u3 3 hexamer in place against the rotation of
the y subunit during rotational catalysis. Based on chemical crosslinking data, it is
currently believed that the a subunit resides to the periphery of the ring of c subunits with
the membrane-spanning domain of the b dimer situated to one side of the a subunit where
it is in close proximity to both the a and c subunits. The b dimer extends out of the
membrane and in a highly elongated conformation reaches to near the top of F1, making
contacts with the a and p subunits along the way and the 6 subunit at its extreme
The stator stalk consists of the b subunit of Fo and the 6 subunit of F1. Although
the primary function of the a subunit of Fo is considered to be a role in proton
translocation along with the clo subunits, it nevertheless plays the part as a stator and will
be discussed in this section. Structurally, the a subunit plays a role in both the formation
of a dynamic interface with the ring of c subunits as well as the formation of a secure
complex with the b dimer. Pursuit of a high-resolution structure for the a subunit remains
a challenge to this day. The high-resolution structure of the region of the a subunit that
forms the interface with the clo subunits is eagerly anticipated since it appears to be the
crucial region for proton translocation.
In regard to the stator stalk, partial structural information has been obtained for the
E. coli b subunit membrane spanning domain and the 6 subunit amino terminus by NMR
studies (136, 137). X-ray crystallography has solved the structure of a model polypeptide
based on the dimerization domain of the b subunit (138). The binding of F1 to the
membrane-bound Fo requires both the 6 and a subunits, suggesting that each of the
subunits are involved with the stalk structures of F1Fo ATP synthase (87, 139, 140). In
fact, the 6 subunit forms an integral part of the peripheral stalk and the a subunit
functions as part of the central stalk (discussed below). The 6 subunit of F1 has been
visualized seated at the very top of the Fl u3 3 hexamer by EM (11). However, recent
evidence has suggested that the 6 subunit may actually be positioned slightly to the side
of F1 in association with only a single oc subunit (Figure 1-2) (100, 141-143). The b
subunit is the primary focus of this dissertation and will be discussed at length later in
The E. coli a subunit is a large, extremely hydrophobic protein encoded by the
uncB gene and consists of 271 amino acids with a molecular weight of 30,276 Da. All
enzymes of the F1Fo ATP synthase family contain an a subunit homolog with strong
primary sequence homology even among evolutionarily diverse species (144). The most
highly conserved region resides in the carboxyl-terminal one-third end, amino acid
residues al90-263. Notably, in the region that is involved in proton translocation, there is a
remarkable conservation of the amino acid residues aleu207, aarg210, aleu211, asn214 and agln252
and an evident conservation of aglu219 and ahis245 at the homologous positions in all a
subunits from different species (144). The aarg210 iS the most strictly conserved among all
species and does not tolerate substitution with any other amino acid (further discussed
As mentioned above, there is no high-resolution structural data for the a subunit.
Also, contradicting models exist concerning the number of transmembrane helices as well
as the orientation in the membrane. Difficulty in studying the structure of the a subunit
arises from its extreme hydrophobicity and the necessity to include the denaturant,
trichloroacetate, in purification procedures. This is compounded by the fact that it cannot
be expressed at high levels in E. coli and is not found in the membrane without the
presence of both the b and c subunits (148-150). Furthermore, the a subunit is known to
be a substrate of the protease FtsH, which will rapidly degrade the subunit if it is not in
its native state (151). It was readily labeled with TID, which is a hydrophobic carbene
generator that is believed to react from the nonpolar region of the lipid bilayer, but its
solubility properties made it unsuitable for analysis as was done with the c and b subunit
(106). Consequently, the amino acids in contact with the lipid phase of the bilayer were
not identified. Due to difficulties in obtaining high-resolution structural data, much of
what is known of the a subunit arises from mutational studies.
Topology. Hydropathy analyses indicated five definite membrane-spanning
regions and one putative membrane span (121, 140, 152, 153). Much of what is known
of the a subunit structure and has come from the analysis of cysteine mutagenesis.
Greater than 50 cysteine substitutions, which resulted in a functional F1Fo ATP synthase,
were used in two kinds of experiments (154-158). Various maleimide derivatives were
used to search for the surface-accessible regions (154, 155). And double cysteine
mutations were used to search for disulfide formation between a-a and a-c (153). The
results supported the model in which the a subunit spans the membrane five times and the
fourth span, which includes aarg210 iS in contact with the second transmembrane ot-helix of
the c subunit (Figure 1-4A). Additionally, residues that were originally thought to be
located in the cytoplasm were not labeled, indicating that the six-membrane span model
was incorrect (132, 159).
The location of the amino-terminus of the a subunit has also been very
controversial. A substantial amount of evidence indicates that the carboxyl terminus
4~ 1 154
Figure 1-4. Controversial models of the a subunit topology. There is no high-resolution
structural data for the a subunit. Mutagenesis, crosslinking and
immunological experiments were used to study the topology. Roman
numerals indicate the number of transmembrane helices. Small numbers
indicate the relative position of the amino acid residue. Several crosslinking
reactions were observed between the fourth helix of the a subunit (IV*) and
the second helix of the c subunit in double cysteine mutants. Contradicting
models exist for the topology of the a subunit. A) Model with five-
transmembrane helices and the amino terminus residing in the periplasm. B)
Model with six transmembrane helices and the amino terminus located in the
resides in the cytoplasm (154, 155, 160). This observation, in combination with the five-
transmembrane helices, indicates that the amino-terminus should reside in the
periplasmic space. Polyclonal antibodies against a peptide model of the extreme
carboxyl-terminus as well as antibodies against epitope tags constructed at the carboxyl
terminus of the a subunit revealed this region to be located in the cytoplasm (160).
Moreover, cysteine substitutions at a266 Of G277 were highly reactive on the cytoplasmic
side of the membrane (154, 155). The orientation of the amino and carboxyl-termini was
studied by gene fusion proteins and peptide-directed antibodies, revealing a cytoplasmic
location of both termini (161, 162). Insertion of epitope tags at various positions also
confirmed the cytoplasmic local of both termini, arguing in favor of the controversial six-
transmembrane model of the a subunit (Figure 1-4B) (160). In the five-transmembrane
model a stretch of about 37 amino acids at the amino-terminus resides in the periplasm
with only one transmembrane helix, approximately up to residue a66, preSent (Figure 1-
4A) (154-156, 158). In the six-transmembrane model the amino-terminus resides in the
cytoplasm with two transmembrane helices present before the first cytoplasmic loop,
which range from approximately residues a33-49 and a54-70 (Figure 1-4B) (160).
A series of a subunit amino-terminal truncations and internal deletions were
constructed and the F1Fo ATP synthase function was tested by growth on a succinate
minimal media. Assembly of intact complexes was tested by membrane-associated
ATPase activity and the presence of the a subunit was analyzed by immunoblot analysis
(163). Four sections were found to be particularly interesting. The first 33 residues at the
amino terminus were shown to be necessary for the insertion of the a subunit into the
membrane. Two internal deletions, from residues a91-99 and al63-177, TOSulted in functional
enzyme complexes, indicating that these regions were not important for function. A
fourth deletion, from residues al20-124 WAS concluded to be important for function, but not
assemble because high levels of a subunit were found in the membrane, but the enzyme
was not functional.
The importance of the carboxyl-terminus was also analyzed by constructing a series
of early termination codons (164). Sequence alignment of the a subunit demonstrates
that many bacterial homologues contain glutamate and histidine residues at the extreme
carboxyl-terminus (glu-glu-his in E. coli). However, truncation of the Einal four residues
had no effect, and truncation of the Einal nine residues were tolerated at 250C, suggesting
that the extreme carboxyl terminus of the a subunit did not significantly contribute to
proton conduction or functional interactions with other subunits.
Proton translocation. The first indication that the a subunit was directly involved
in proton translocation appeared nearly two decades ago when mutations constructed in
the a subunit (aser206-leu and ahis245-tyr) WeTO found to affect Fo-mediated proton pumping
without influencing F1Fo ATP synthase assembly (165). Since then, not including the
cysteine mutations described above, more than 75 missense mutations have been
constructed and analyzed in or near the conserved regions of the a subunit to chart the
amino acids involved in proton translocation. In general, mutation of a conserved amino
acid residue impaired Fo-mediated proton translocation, but the severity of the defects
The only F1Fo ATP synthase a subunit residue that is strictly conserved amongst all
species, from bacteria to humans, and cannot endure any amino acid substitution, whether
basic, acidic or nonpolar, was aarg210 (145-147). Mutations at this site abolished both
ATP-driven proton pumping and passive Fo-mediated proton translocation. Growth on
succinate minimal media indicated no ATP synthesis by the mutants. The observed
effects were shown not to be due to failure of F1Fo ATP synthase to assemble because
treatment with the detergent lauryldimethylamine oxide (LDAO) released Fl from the
prepared membranes and revealed abundant ATP hydrolysis activity. The presence of
assembled F1Fo ATP synthase complexes incorporated with an aarg210 mutant was later
directly confirmed by Dr. James Gardner (167). Substitution with an alanine allowed
passive Fo-mediated proton translocation indicating that the proton channel was intact
and suggested that the aarg210 iS not obligatorily protonated or deprotonated during proton
conduction (168). A second site suppressor mutation, agln252 arg, which partially
compensated for the aarg210 gln mutation, was identified, and suggested to be in close
proximity to each other with residence on the transmembrane helix 5 and 4, respectively,
in the five-transmembrane model (121). The a210 TOSidue is thought to have a direct role
in proton translocation. The orientation of the a subunit' s fourth transmembrane helix
had been determined relative to the orientation of the c subunit' s second transmembrane
co-helix by crosslinking double cysteine mutants (157). Crosslinking data has positioned
a214 in ClOse proximity to c62 and c65, and a211 ClOse to c69 (157). This places the putative
fourth helix of the a subunit in contact with the second helix of the c subunit. Models
have the a210 TOSidue positioned near the center of the fourth helix at a level in the lipid
bilayer very close to the essential casp61 TOSidue (14). Whether a210 iS directly
protonated/deprotonated or controls protonation of the casp61 TOSidue remain unanswered
(169). Insight from a high resolution structure of an intact Fo is greatly desired and
would provide extremely valuable answers to many of the unsolved questions.
Single mutations at residues a218, a219 Of G245 WeTO Shown to have a considerable
impact on Fo-mediated proton conduction (144, 146, 170). When comparing amino acid
sequences of various mitochondria, chloroplast and bacteria, there appears to be an
instance of evolutionary covariation with these three amino acids (144). This suggests
that when a mutation occurred in one of the three residues, it was accompanied by a
second mutation to compensate for any loss in activity. This would cause the two
residues to pass through evolution as a hereditary unit. Based on this observation, double
mutants were constructed in the E. coli a subunit to imitate other lines of evolution (144,
171). Every double mutant studied resulted in functional F1Fo ATP synthase complexes
with considerably more activity than any of the single residue mutants. Due to the
functional relationship, it is possible that these three amino acids are in close proximity to
A few other strongly conserved amino acid residues located on the fourth and fifth
transmembrane helices are worth mentioning. Residues aasp214 and agln252 were both
strongly conserve but found nonessential, with the effects of mutations at these residues
varied widely (146, 170, 172, 173). Models of the a subunit have these residues lining a
water-filled proton channel. Recently, the aqueous accessibility of residues along
transmembrane helices 2 and 5 has been shown to extend to both sides of the membrane
(174). Also, a mutation at residue 217, aala217,ar,, blocked proton conduction and
inhibited F1Fo ATP synthase activity (167).
The 8 subunit
The E. coli 6 subunit is one of the F1 subunits. It is discussed here because it is an
essential part of the F1Fo ATP synthase stator stalk. The simplest stator stalks occur in
nonphotosynthetic prokaryotes and consist of a dimer of the Fo b subunits (discussed in
the following section) and a single Fl 8 subunit. The 6 subunit displays a very low level
of conservation across various species. It is a globular protein encoded by the uncH gene
that consists of 177 residues with a molecular weight of 19,332 Da. It plays essential
roles in both the binding of Fl to Fo as well as coupling of the catalytic activities of Fl
and Fo (139, 175-180). Circular dichroism (CD) spectroscopy and sedimentation analysis
studies performed on the 6 subunit suggested a highly helical and elongated conformation
Structure of the 8 subunit. A partial high-resolution structure of the 6 subunit has
been solved by NMR (137). During the purification procedure, a truncated form of the 6
subunit was produced by a bacterial protease. It was revealed to be the amino-terminal
134 amino acids (81-134) by mass spectroscopy and N-terminal sequencing. The same
sized 6 subunit fragment was often seen in Fl preparations and could be produced in
isolated E. coli Fl by treatment with trypsin without liberating the 61-134 frOm the Fl
complex (89). Furthermore, purified 81-134 Stably binds to 6-free Fl preparations. The
high affinity of the 61-134 Subunit for Fl indicated that the conformation of the 6 fragment
was preserved during the purification procedure. NMR was performed on both the 61-134
fragment and the intact 8 subunit; however, the quality of data for the intact subunit was
not sufficient enough for structural analysis due to its propensity to aggregate at high
concentrations. To date, the carboxyl-terminal 43 amino acid residues (6135-177) Of the 6
subunit is the only portion of the Fl sector not known at the atomic level.
The amino-terminal 105 residues of the 6 subunit formed a dense globular domain,
while the region from residues 106-134 was mostly disordered with the exception of one
a-helix (137). The amino-terminal domain, 61-105, consisted of a six a-helix bundle with
the dimensions 45 x 20 x 30 A+. Helices 1 (64-20) and 2 (624-38), and helices 5 (670-81) and 6
(688-104) Organized into V-shapes that intercalated to form a core. Helices 3 (641-47) and 4
(653-64) were packed compactly against this four-helix core. Following this globularly
packed domain there was a loop region followed by a seventh a-helix (6118-129).
Comparison of the structural data for the intact 8 subunit against that of the 61-134
fragment illustrated the same structure for residues 1-104, but the spectral shift of
residues 105-134 was very different. It was possible that the carboxyl-terminal 42
residues missing from the 61-134 fragment affects this region of the 6 subunit.
8 subunit topology. Taken together with biochemical and immunological data, the
structure revealed by NMR revealed that the 6 subunit consists of two domains, an amino
terminal domain, 61-104, and a carboxyl-terminal domain, 6105-177. Under oxidizing
conditions, two native cysteines present in the amino- and carboxyl-terminal domains of
the intact 8 subunit, Scys64 and Scysl40, TOSpectively, formed a disulfide bond. Furthermore,
NMR data indicated some NOE's between the carboxyl-terminal a-helix and the amino-
terminal domain. The data indicated that there is probably a close interaction between
the amino- and carboxyl-terminal domains of the intact 8 subunit.
Proteinase accessibility and immunological analysis were used to examine the
topology of the 6 subunit (89). The 6 subunit was susceptible to trypsin digestion at the
carboxyl-terminal 20 residues in isolated Fl, but not in intact F1Fo ATP synthase,
indicating a protection of the amino-terminal region by Fl. Deletion analysis of the
carboxyl-terminal region also implied the importance of the 6 subunit in binding Fl to Fo.
Taken together, these observations suggested that the amino-terminal domain is
predominantly involved in the binding of the 6 subunit to Fl and the carboxyl-terminal
domain is involved in binding to Fo.
The location of the 6 subunit has had a history of being very controversial. Prior to
the high resolution structure obtained by Abrahams et al. (1994), the b and 6 subunits
were expected to form part of the central stalk of the F1Fo ATP synthase enzyme, which
is now known to consist of the y and a subunits (20). Due to the dimensions, it seemed
unlikely that the b and y subunits could fit as part of the central stalk, which implied that
they must form a separate connection between Fl and Fo. Improving EM technology did
not allow visualization of the second stalk structure at the periphery of the F1Fo ATP
synthase complex until many years later (27, 28). Prior to visualization by EM, several
early crosslinking studies had been reported in the quest to find the location of the
6 subunit binding on Fl, finding it to be on the a subunit (89, 181-184). Notably,
crosslinking the 6 subunit to the a subunit did not have a great impact on F1Fo ATP
synthase function, as would be expected if the 6 subunit formed part of the stator stalk
(185). High-resolution structure of the Fl u3 3 hexamer with a partial structure of the y
subunit had revealed a dimple in the top of the hexamer approximately 15 A+ deep that
was adj acent to the core space where the amino- and carboxyl-terminal a-helices of the y
subunit resided (20). EM studies had revealed a "cap" structure at the very top of F1 in
both E. coli and mitochondrial complexes (27, 29, 186). It was thus believed that the 6
subunit resided in the dimple of F1 as the cap seen in the EM structures (187). This
possibility was refuted when Prescott et al. demonstrated the ability to stably incorporate
the green fluorescent protein (GFP), via varying length peptide linkers (0, 4 or 27 amino
acids), to the carboxyl-terminus of the y subunit without interrupting function of the
enzyme complex. GFP forms a rigid, stable structure with the dimensions 24 A+ wide and
48 A+ high (188). This study indicated that the putative cap structure could not possibly
occupy the entire dimple atop Fl. More recent evidence has suggested that the 6 subunit
may actually be positioned slightly to the side of F1 in association with only a single a
subunit (Figure 1-2) (100, 141, 142, 189).
The b subunit is required for the normal assembly and function of F1Fo ATP
synthase (190). The E. coli F1Fo ATP synthase has two identical b subunits, which form
a homodimer, that are the product of a single gene (Figure 1-2). It is an elongated
amphipathic polypeptide that crosses the membrane one time at its amino-terminus and
has an extensive hydrophilic carboxyl-terminal domain. This pattern is characteristic of
b-type subunits of ATP synthases, although the mitochondrial b has two consecutive
membrane-spanning segments at the amino-terminus (191). Most ATP synthase b-type
subunits consist of between 150 and 170 amino acid residues. The E. coli b subunit,
encoded by the uncF gene, consists of 156 amino acid residues and has a deduced
molecular weight of 17, 265 Da (Figure 1-5).
Domains and Structure. Currently there is no high-resolution structure of the
entire b subunit. Several factors probably contribute to the difficulty of structural
analysis. The b dimer is a thin, highly extended, mostly a-helical structure, its
dimerization is comparatively weak and reversible (192), and it displays evidence of
flexibility (193-195). This has led to alternative low-resolution approaches to study the
structure of the b dimer such as circular dichroism (CD) spectroscopy, deletion analysis,
74=, 75= 78 227= 228 a,
t 200 bend a a
++ A* A* A f 40
MNLNATILGQ AIAFVLFVLF CMKYVWVPPLM AAlEKRQKEl
At* .. X 8
ADGLASAERA HKDLDLAKAS AT
QVAILAVA GAEKllERSV DEAANSDIVD KLVAEL
Figure 1-5. Amino acid sequence of the E. coli F1Fo ATP synthase b subunit. The E. coli
b subunit is a 156 residue amphipathic polypeptide. The amino acid sequence
and the four domains are shown. The transmembrane domain (bl-22), tether
domain (b24-60), dimerization domain (b63-122) and the 6-binding domain (bl23-
156) are Shown in blue, orange, green and red, respectively. The large purple
stars indicate residues capable of forming high yields of b-b crosslinks upon
cysteine substitution. The smaller purple stars indicate residues found to form
low-yields of crosslinks. The arrows indicate positions crosslinked to other
subunits of ATP synthase. High-resolution structures based on model
polypeptides consisting of bl-34 and b62-122 (underlined residues) have been
solved by NMR and crystallography, respectively. NMR analysis of residues
1-34 has revealed a a-helical structure with a rigid 200 bend at positions 23-
26. X-ray crystallography revealed a highly a-helical structure with modeled
into a right-handed coiled coil.
analytical ultracentrifugation, chemical crosslinking, and the analysis of tendencies for
disulfide bond formation. CD spectroscopy analysis has predicted the secondary
structure of the b subunit to be approximately 80% ot-helical with about 14% p-turn
Although there is no high-resolution structure of the intact Fo sector, an abundance
of evidence suggests the necessity of the b subunit to exist in the dimeric state. The
hydrophilic region of b, consisting of residues b24-156 (alSo known as bsol), has been
expressed and shown to form highly extended dimers capable of binding to F1-ATPase in
solution (197). Sedimentation equilibrium ultracentrifugation gives a molecular weight
value of about 30,000 Da for bsoi, consistent with a dimer of two 15,000 Da bsol
monomers (13). The existence of the dimeric state of the b subunits was confirmed by
covalently cross-linking the two b subunits in the complex and verifying the activity of
the enzyme (198). Furthermore, the ability of b to bind to Fl was discovered to be
directly proportional to the ability of b to form dimers, suggesting the necessity of the b
dimer formation before the binding of Fl to the complex (199).
The dimerization of the b subunit has been shown to be relatively weak and
reversible. The monomeric and dimeric forms of bsol were shown to exist in a dynamic
equilibrium and the dimer was converted to the monomeric state at 400C (192). This
same melting characteristic was observed with CD spectroscopy (200). Furthermore, the
similar traits were observed in photosynthetic organisms, which encode two different b-
type subunits, b and b '(13). When the cytoplasmic regions of the b and b subunitss from
the cyanobacterium Synechocystis were expressed individually, the polypeptides were
found to only exist in the monomeric state. However, when they were mixed together,
the formation of the dimers was observed by chemical crosslinking and sedimentation
equilibrium ultracentrifugation (13). Also, the dimers were observed to melt at 400C as
was the case with the E. coli b subunit. The striking similarity between the E. coli b
subunit and the photosynthetic organism b and b subunits indicate that the former is a
good model in which to study the b subunit.
Cross-linking and deletion analysis has led to the development of a four-domain
model of the E. coli F1Fo ATP synthase b subunit (Figures 1-5 and 1-6) (12). Amino acid
FDB 1 adinn
Dim eriz ation
Figure 1-6. Gross structure of the E. coli F1Fo ATP synthase and the domains of the b
subunit. The b subunit domains were described by Dunn et al. (12) The
membrane-spanning domain roughly corresponds to amino acid residues 1-22,
the tether domain is approximately residues 24-60, the dimerization domain is
considered to be residues 60-122, and the F1-binding domain is roughly
residues bl-22 COrresponds to the hydrophilic membrane spanning domain. Residues b24-60
roughly corresponds to the tether domain, which is the portion of the peripheral stalk
often observed in electron micrographs. The dimerization domain, approximately b60-122,
is required for the dimerization of the two b subunits. And finally, the Fl binding
domain, roughly amino acid residues bl23-156, iS required for the binding of F1 to Fo.
The amino-terminal membrane spanning domain, bl-22, forms a single
transmembrane span while the large remainder is a polar hydrophilic domain which
extends above the cytoplasmic leaflet of the lipid bilayer and reaches towards the top of
Fl (Figure 1-6) (13, 191). A wealth of evidence has suggested this proposed topology for
the b subunit. The amino terminal region was uniformly vulnerable to chemical
modification by the a nonpolar photoreactive reagent, TID, which is a hydrophobic
carbene generator that is believed to react from the nonpolar region of the lipid bilayer,
indicating that this region was in fact in the hydrophobic phase of the bilayer (106). This
observation was consistent with other labeling procedures including the labeling of boys21
by hydrophobic nitrenes (201) or the hydrophobic maleimide N-(7-dimethylamino-4-
methyl-coumarinyl)-maleimide (DACM) (148). Modification of boys21 interfered with
intersubunit interactions within Fo. Furthermore, reconstitution of the Fo subunits upon
labeling the b subunit with DACM resulted in reduced proton translocation as well as Fl
binding affinities (148). TID failed to label the region basn2-gln10, indicating that the first
few residues at the amino-terminus protrude into the periplasm (106)
Despite attempts by several laboratories, there is presently no high-resolution
structure of the entire b subunit. Therefore, model polypeptides have been constructed in
order to elucidate the structure of the b subunit by domain. A model polypeptide
comparable to residues bl-34, which contains the membrane-spanning domain, dissolved
in a 4:4: 1 v/v mixture of chloroform/methanol/water previously used for solving the
structure of subunit c, has been solved by NMR (136). The data revealed an a-helical
monomeric structure with a 20o bend at residues 23-26 (KYVW) (Figure 1-5). The
hydrophobic residues, b4-22, formed an a-helix, followed by the 200 bend, and then
resumed with a-helical structure from residues b27-34. The bend was proposed to be
positioned as the b subunit exits the membrane. A series of cysteine substitutions
resulted in a high yield of crosslinks formed at residues b2, b6, and blo (Figure 1-5). A
lower yield of crosslinks were observed to form at residues b3, bs, b9 and bll (Figure 1-5).
No crosslinks were observed when cysteines were substituted for residues 12-21. The
observance of continuous crosslinks between residues b6-11 WeTO Suggested to indicate a
dynamic interaction between the contacting faces of the two b subunits in this region
(132). These observations led to a dimeric model in which the extreme amino-termini of
the b subunits crossed each other in close proximity at an angle of about 350 in the region
of residues b4-11, and then the two b subunits angled apart as they traverse the membrane
towards the cytoplasmic side (Figure 1-5) (136). The region of the 200 bend, b23-26, WAS
suggested to change the direction of the second a-helix, b27-33, Such that it would extend
into the cytoplasm at an angle perpendicular to the plan of the membrane. This model
was confirmed by a systematic mutational analysis of the membrane-spanning domain
performed by Hardy et al. (202).
The tether domain of the b subunit, roughly b24-60, iS the least defined part of the
subunit domains from a structural point of view. It corresponds to the portion of the
peripheral stalk often seen in electron micrographs and is called "tether" simply because
it is the section of the b subunit that links the more defined membrane-spanning and
dimerization domains Figure 1-6). There is no high resolution structure for this region of
the b subunit. The NMR structure of bl-34, described above extended slightly within this
domain, revealing an a-helix at least up through residue b34 .(136) Also, a heptad repeat,
extending from just outside of the membrane and continuing without interruption up to
residue bala79 Suggests the structure to be a coiled coil (197, 203). Though crosslinking
studies showed that the tether domains of the two b subunits are in parallel and in close
proximity, this domain contributes little to the stability of the dimerization of the b
Deletion constructs analyzed by sedimentation equilibrium experiments suggested
that a form of the b subunit truncated in each end, b53-122, WAS capable of forming dimers
with an efficiency close to the complete cytoplasmic domain, b24-156, indicating that the
most pertinent intersubunit contacts of the b subunit was located within this central
region, referred to as the dimerization domain (192). More recently, the dimerization
domain has been refined to residues b63-122; however, the amino-terminal boundary is
likely to decrease even further due to the observation that deletion of residues b54-64 Of
b65-75 TOSulted in intact and functional F1Fo ATP synthase complexes (Figure 1-6) (194).
A crystal structure of a monomeric dimerization domain, based on a model
polypeptide consisting of residues b62-122, has recently been solved and refined to 1.55 A+
(138). Based on an undecad repeat and crosslinking data, Dunn and coworkers have
constructed a model in which the two a-helices of the b62-122 TegiOn formed a coiled-coil
with a right-handed superhelical twist. A number of previous studies supported this
coiled-coil arrangement. First, the shape of the b53-122 pOlypeptide was consistent with a
coiled-coil of similar length as determined from its frictional coefficient (1.60) in an
ultracentrifuge and from NMR relaxation parameters (192). Secondly, small-angle X-ray
scattering by b52-122 in Solution specified a maximum length to be about 95 A+, consistent
with the expected coiled-coil length (13). Thirdly, CD spectroscopy indicated that this
polypeptide was 100% ot-helical and the similar intensities of the minima suggested the
helices to be arranged in a coiled-coil (204). Fourthly, cysteine substitution and
crosslinking studies suggested a periodicity consistent with a parallel coiled-coil (Figure
1-5) (204). Finally, b subunit sequence analysis of E coli and other prokaryotes revealed
a conservation of an undecad pattern, which is a distinctive characteristic of a right-
handed coiled coil. Mutation of amino acid residue barg-83, which interrupts the undecad
repeat, markedly stabilized the dimer, as expected for the proposed two-stranded, right-
handed coiled-coil structure.
The carboxyl-terminal F1-binding domain, bl24-156, alSo referred to as the 6-binding
domain, has a more globular conformation and is required for the binding of F1 to Fo
(205, 206). Work accomplished by Futai and coworkers two decades ago revealed that
truncation of the extreme carboxyl-terminus of the F1-binding domain by only a few
amino acids resulted in assembly defects in F1Fo ATP synthase (206). Subsequently,
work performed by Dunn and coworkers demonstrated that the final two to four amino
acids of the b subunit were necessary for binding the 6 subunit of Fl (205). An addition
of a cysteine at the carboxyl-terminus was chemically crosslinked to a cysteine
introduced at Siss.
A close association of the two b subunits in the F1-binding domain was indicated
by crosslinks formed between cysteines individually substituted at positions bl24, bl25,
bl26, bl27, bl28, bl29, bl30, bl31, bl32, bl39, bl44, bl46, Or bl56 (Figure 1-5) (198, 200, 207).
Hydrodynamic evidence favors a folded structure for this domain of the b subunit as
opposed to the highly elongated structure of the remainder of the b subunit. Also, several
studies have shown that either a balal28-glu mutation, deletion of the last four residues, or
cold temperature dramatically decreased the sedimentation coefficient, by 23%,
suggesting that the F1-binding domain underwent a conformational change from a
globular structure to a less folded more extended conformation (192, 205, 208). The
mutation, balal28-glu, may have caused an electrostatic repulsion that would cause the two
b subunits to push apart. The carboxyl-terminal residues may form an amphipathic helix,
so the deletion would have disrupted essential interactions. And cold temperatures have
been shown to weaken hydrophobic interactions in proteins, suggesting the importance of
the hydrophobic amino acids in the folding of this domain (13). Dunn et al. suggested
that these observations implied that the carboxyl-terminus of the b subunit has a weakly
folded structure in which the hydrophobic amino acids are arranged to impart structural
stability and create hydrophobic patches on the surface (13). The folded conformation
appears to be required for the exposed hydrophobic patches to interact with Fl.
Mutagenesis. Several mutant searches and site-directed mutagenesis studies have
been performed in the membrane-spanning domain of the b subunit. However, only a
single mutation, bgly9,asp, lOcated near the periplasmic side of the lipid bilayer, resulted in
a defective proton pore in an intact F1Fo ATP synthase complex (209). Second site
suppressors of this mutation have been found in the a(apro240-ala Of Opro240-leu) and c
(cala62-ser) Subunits that partially repaired the defect, indicating an interaction between the
b subunit and both the a and c subunits (210, 211). The membrane-spanning domain
contains one charged residue, blys23, but mutations generated at this site did not influence
proton translocation, suggesting that this domain of the b subunit did not have a direct
function in proton conduction, although it was required for the maintenance of a
functional Fo complex (212). A systematic mutational study of the membrane-spanning
domain conducted by Hardy et al. was described above and a triple mutant, bN2A,T6A,Q10A,
is described in Chapter 5 of this dissertation (202).
A couple notable characteristics can be attributed to the tether region of the b
subunit. Relatively large deletions and insertions of up to 11 and 14 amino acids,
respectively, were accommodated in this region and the altered b subunits still assembled
into fully functional F1Fo ATP synthase complexes (193, 194). In fact, decreases
observed in enzyme activity paralleled the decrease of b subunit found in the membrane,
suggesting that the alterations affected assembly of the enzyme, but not the function.
Assuming co-helical structure, an 11 amino acid deletion would shorten the b subunit
tether region by approximately 16.5 A+ and a duplication of 14 amino acids would
increase the length by about 21 A+. This implies that the b subunit is highly flexible and
the altered b subunits may compensate for the lost or gained distance via that flexibility.
The fact that the peripheral stalk must extend from within the membrane to up near the
top of Fl suggests that some part of the stalk must be flexible enough to stretch or
straighten in the shortened b subunits, or in the case of the lengthened b subunits, bend to
take up the slack. Prior to these observations, the b subunit was often viewed as a rigid,
rod-like structural feature during rotational catalysis. Also, though the b subunit is the
least conserved subunit of F1Fo ATP synthase, gaps are rarely found in sequence
alignments of numerous organisms (213, 214). Therefore, the ability to manipulate the
Figure 1-7. Model for F1Fo ATP synthase peripheral stalk orientation dependent upon the
direction of rotation during ATP synthesis or hydrolysis. The subunits are
color coded as follows: ot, light blue; P, grey; y, dark blue; E, green; 6, orange;
a, yellow; b, red; c, cyan. The panel on the left indicates the orientation of the
b2 dimer if the enzyme is actively synthesizing ATP. The panel on the right
represents the position of the b2 dimer during ATP hydrolysis. The arrows
indicate the direction of rotation of the rotor stalk subunits (ysclo). The red
cylinders indicate regions of the b subunits for which there is no high-
length of the peripheral stalk was an unexpected surprise to the field. The length of the
wild-type b subunit is probably the optimum length for assembly of F1Fo ATP synthase.
The apparent flexibility has been proposed to help alleviate torsional strain brought about
by rotational catalysis (193). Another hypothesis concerning the flexibility of the tether
domain is that this region of the b2 dimer serves as a hinge, allowing reorientation of the
stator depending on the direction of rotation as the enzyme carries out ATP synthesis or
hydrolysis (Figure 1-7) (195, 202).
Another important feature of the tether domain is the evolutionarily conserved
barg36, that has been implicated in a structural role influencing proton conduction through
Fo. Mutational studies at this amino acid residue led to numerous defects from failure to
assemble or function to uncoupling phenotypes (215). Amino acid substitutions, barg26-ile
or barg26-glu, TOSulted in assembled F1Fo ATP synthase complexes that displayed defects
in Fo-mediated proton translocation or a disruption of coupling activity, respectively.
Substitution with a cysteine at this residue led to a crosslink product with the a subunit of
Fo (216, 217). The close proximity of this conserved residue to the a subunit indicates
that it may play a role in aligning the proton exit channel.
Protein-protein interactions between the two b subunit dimerization domains have
been shown to be essential for forming the peripheral stalk (13). Mutations at a
conserved residue, bala79, TOSulted in maj or F1Fo ATP synthase assembly defects (203,
218). The bala79 mutations were modeled in the bsol polypeptide to investigate the affects
of the mutations. The model polypeptides were shown to retain co-helical structure, but
chemical crosslinking and sedimentation experiments suggested that the bala79 mutants
were incapable of forming dimers (199).
In the F1-binding domain, a mutation was found, balal28-asp, that had little effect on
the dimerization of the b subunits but led to an assembly defect of F1Fo ATP synthase
(208). However, the mutant was found to have a reduced tendency to interact with Fl and
sedimentation equilibrium ultracentrifugation experiments revealed a 12% decrease in the
sedimentation coefficient, indicating a structural perturbation (discussed above). The
studies suggested that the balal28 TOSidue was not important in b subunit dimerization, but
it had an important structural role in the F1-binding domain.
Intersubunit interactions. The formation of disulfide bonds between cysteine
residues introduced in the membrane-spanning domain of the b subunit and the a subunit
as well as second site suppressors of the bgly9,asp found in the a subunit (discussed above)
strongly suggests an interaction between the two stator subunits (209-211). The bgly9-asp
mutation was also partially suppressed by a mutation in the c subunit (discussed above),
but it is not known whether it is due to a direct interaction between the b and c subunits
or if the suppression is mediated through the a subunit.
The b subunit has also been shown to interact with the ot, P and 6 subunits of F1
(198, 216). The interaction of the Fo b subunit with the Fl 8 subunit has been well
documented (200, 205, 219, 220). The interaction is mediated by the carboxyl-termini of
both subunits and is essential for the binding of F1 and Fo. The critical role of the 6
subunit mediating the interaction between the b subunit and the bulk of F1 has been
demonstrated by the inability of 6-depleted Fl to bind to Fo (219). Crosslinking the b and
6 subunits via introduced cysteines did not affect F1Fo ATP synthase activity, which was
consistent with the proposed role of the b28 peripheral stalk as a stator and demonstrated
that the b-6 interaction need not be dynamic (221). Though it is believed that the binding
of Fl to Fo heavily relies on the b-6 interaction, evidence of other subunit contacts may
also influence binding. Furthermore, the binding of b to 8 was shown to be relatively
weak by analytical ultracentrifugation, indicating that other subunit contacts may
contribute to binding (220). Many crosslink formations were found when cysteines were
introduced into the b and a or P subunits (Figure 1-5). A cysteine introduced to the
carboxyl-terminus of the b subunit has been shown to crosslink to acvs90 (198). Also,
cysteines positioned at b92 Or bl09 formed crosslinked products with the a subunit or both
the a and p subunits, respectively (216). These results confirm that the b subunit is
proximal to the u3 3 hexamer, but a direct interaction has not been confirmed.
Stator stalk function. The necessity of a stator in F1Fo ATP synthase was
recognized after the realization that rotation was a fundamental feature of ATP catalysis.
The current view of the peripheral stalk is primarily that of the stator which forms a
connection the a subunit and the u3 3 hexamer, holding these subunits in place against
the rotation of the rotor subunits, cloys. The idea of a flexible stator stalk has led to other
proposed features of the b subunit. The apparent flexibility of the b subunit has been
suggested to transiently store energy during rotational catalysis which could be
potentially be expended to force the conformational change that allows the release of
ATP(222). Another model describes the flexibility of the tether region as a hinge that
could reorient the b subunit when switching between ATP synthesis and hydrolysis
(discussed above) (Figure 1-7) (202). However, there is no direct evidence of whether
the b subunit is actually acting in a flexible manner. Finally, the b subunit has recently
been suggested to influence the nucleotide binding sites in the P subunits (223). In these
experiments, a spin label was incorporated at residue P331. Upon addition of bsoi, the
spectrum of this spin label was observed to change in a way that implied that the catalytic
sites were in a more open conformation. These results indicate that the current view of
the stator stalk as a structural feature may soon be revised such that the b subunit has a
more direct role during rotational catalysis.
The E. coli ATP synthase complex has, by far, the simplest architecture of all the
F1Fo ATP synthase enzyme complexes. It is composed of twenty-two polypeptides of
eight different types with the stoichiometry u3 3y~sab2010 (Figure 1-2, Table 1-1) (112,
224). The u3 3YiS Subunits comprise the Fl sector and the ab2C10 Subunits comprise the
Fo sector. In the E. coli enzyme, all eight subunits are necessary for the function of F1Fo
ATP synthase (225, 226). Chloroplasts also have a relatively simple architecture with the
exception that they have nine different subunits (227) due to the fact that the two b
subunits are products from two different genes and are not identical (Figure 1-8). In
contrast, the F1Fo ATP synthase from mammalian mitochondria is composed of at least
thirty-one polypeptides of sixteen different types with the Fl stoichiometry of u3 3Y68
and a much more complex Fo consisting of a, b2, C10-14, d, e,J; g, (F6)2, A6L, OSCP, IFI
(7, 228-230). Yeast mitochondria F1Fo ATP synthase has an extra three subunits
compared to the mammalian enzyme, stflp, i, and k.
The F1 sector. In the Fl sectors, homologues of the E. coli F1Fo ATP synthase
have been identified for the a, P and y subunits, based on amino acid sequence
homology, in the chloroplast and mitochondrial enzymes (30). Based on the primary
sequences, the highest conserved subunit from the E. coli F1Fo ATP synthase is the P
subunit with approximately 70% homology with the chloroplast and mitochondria
Table 1-1. F1Fo ATP synthase subunit equivalency
Bacteria Chloroplast Yeast Bovine Function
a a a a catalytic site
P P P p catalytic site
y y y rotor
6 6 OSCP OSCP stator
a E 8 8 rotor
-~ E stabilization?
a a (or IV) a (or 6) a poton channel, stator
b b (or I)b (or 4) b stator
Sb '(or II) 9 -stator
c c (or III) c (or 8) c poton channel, rotor
-~~ d'a stator?
8 A6L stator?
e e ?
h F6 stator
inh1p IF1 inhibitor
equivalents (31). The a subunits exhibit roughly 50% homology (31). The nucleotide
binding regions of these two subunits also have sequence homologies with other proteins
that bind nucleotide or phosphate, including the E. coli secA protein, N-ethylmaleimide
sensitive fusion protein, herpes simplex virus UL15, Ca2+-ATPase, H /K+ ATPase and
Na /K+ ATPase (34-37). Furthermore, the nucleotide binding motif, GXXXXGKT/S,
which was first identified in the a and P sequences ofF1, has been found to be conserved
in the high-resolution structures of other proteins including p21,as, adenylate kinase,
RecA, elongation factor Tu, and transducin-a (20, 38-42).
Interestingly, the y subunit in the chloroplast F1Fo ATP synthase complex contains
an insert of about 35 amino acids that is not present in the mitochondrial or
nonphotosynthetic eubacteria (231). This loop contained two cysteine residues that were
found to be reduced in the active enzyme complex during photosynthesis and oxidized to
a disulfide bond in the inactive enzyme while in the dark (232). The E. coli s subunit is
unfortunately known as the 6 subunit in the mitochondrial enzyme (233) and also shares
primary sequence homology with the mitochondrial IF1 inhibitor protein. The E. coli s
and bovine 6 subunits share 60% sequence identity, which had suggested that they are
functionally equivalent (191, 234). The availability of high resolution structures for these
subunits has revealed that they are strikingly similar. Superimposition yields a 1.64 A+
rms deviation (48). The bacterial a subunit has been suggested to be an inhibitor of ATP
hydrolysis, undergoing large, ratchet-like conformational changes to selectively switch
off ATP hydrolysis (102). In mitochondria, this inhibitor action of the bacterial a subunit
is ascribed to the IF1 protein of the Fo sector. It was suggested that the bacterial a subunit
was separated into the two polypeptides in the mitochondrial enzyme complex, E and IF1.
Finally, the bacterial and chloroplast 8 subunits ofF1 share a significant sequence
homology (234). The 6 subunit's equivalent in the mitochondrial F1Fo ATP synthase is
known as oligomycin-sensitivity-conferring protein (OSCP) (23 5-23 7). The carboxyl-
terminal region of the mitochondrial b subunit was been demonstrated to bind to the
OSCP subunit (E. coli 6 subunit) through subunit interactions (Collinson et al., 1994) and
chemical crosslinking analysis (Soubannier et al., 1999).
The Fo sector. The Fo sectors of the F1Fo ATP synthase family is by far more
diverse than the Fl sector with an additional eight different subunit types in mammalian
and an extra ten different subunits in yeast mitochondria (Table 1-1). The E. coli a and c
subunits are respectively equivalent to the chloroplast IV and III subunits and the
mitochondrial ATPase-6 and ATPase-8 subunits. In every case, both subunits are
hydrophobic proteins required for proton translocation (120, 140). Unlike the a and c
subunits, the E. coli b subunit does not have any obvious homologues in chloroplasts or
cyanobacteria F1Fo ATP synthases. However, both of them have two distinct subunits
with similar hydrophobic and hydrophilic residue distribution (238). These subunits are
referred to as subunits b and b 'in cyanobacteria and subunits I and II in chloroplasts. It
is believed that only one of each of these subunits is present in the F1Fo ATP synthase
complex, incorporating into the enzyme as a b-b heterodimer as opposed to the b subunit
homodimer present in E. coli. No obvious homologue of the E. coli b subunit has been
found in the mitochondrial F1Fo ATP synthase, even upon analysis of sequence, function
or three-dimensional structure (239). However, hydropathy plot analysis does indicate
that the mammalian mitochondrial b subunit that may be analogous to the E. coli b
subunit (240, 241). Proteolysis studies and crosslinking data supported the location of
the mitochondria b subunit at a position analogous to the E. coli b subunit (242, 243). At
least three other subunits may play the role of the E. coli b dimer including subunit 8 (or
A6L), d' and F6. The mitochondrial b subunit is believed to have two a-helical
transmembrane spans at its amino-terminus arranged in an antiparallel configuration.
The extreme amino-terminus of the mitochondrial b subunit is thought to begin on the
cytoplasmic side of the membrane, traverses the membrane to the periplasmic leaflet of
the membrane as an a-helix, then turn back and traverses the membrane again as it exits
the membrane in the cytoplasm and reaches towards the top of F1 (Figure 1-8). The
remainder of what would be equivalent to the E. coli b dimer is highly speculative,
though the overall shape and characteristics of the b8dF6 subunits favor this explanation
(Figure 1-8). In this model, the mitochondrial subunit 8 contributes to a third membrane
spanning region, and a combination of subunit d and subunit F6 forms the hydrophilic
The Fo subunit known as the a subunit in mitochondrial F1Fo ATP synthase has no
counterpart in the bacterial or chloroplast enzymes. It is a small polypeptide (50 amino
acids) folded into a helix-loop-helix. It is believed to play a role in the stabilization of the
central stalk and its absence in the bacterial and chloroplast enzymes may explain why
Figure 1-8. Speculative models for the b-like subunits. Shown are models for the
bacterial, chloroplast and mitochondrial b subunits. The membrane-spanning
regions are indicated by the black lines. An abundance of evidence supports
the parallel arrangement of the bacterial b2 homodimer and the chloroplast bb '
heterodimer. The mitochondrial analogue of the b subunit is believed to
consist of up to four polypeptides. The model shown for the mitochondrial
b8dF6 Structure is highly speculative.
the bacterial a subunit of the Fl sector (equivalent to the 6 subunit in mitochondria) easily
dissociates from the Fl complex whereas this has not been observed for the mitochondrial
enzyme. Mitochondrial Fo consists of several additional subunits not found in the E. coli
or chloroplast enzymes including the e, fand g subunits as well as an extra three in yeast,
stflp, j, and k subunits.
E. coli F1Fo ATP synthase as a model. Initial studies of the F1Fo ATP synthase
complex were achieved with enzymes isolated from mitochondria or chloroplast.
Although the bacterial, chloroplast and mitochondrial enzymes differ in oligomeric
complexity, the enzymes show acceptable overall structural resemblance and primary
sequence homology that it is widely accepted that the mechanism of action is the same in
all organisms (240, 244). Therefore, studies using the bacterial model became widely
accepted since it was more versatile and offered a large range of research that could not
be readily undertaken with the more complex organisms. Other advantages include ease
of genetic techniques, the ability of bacteria to grow via glycolysis, which allowed
characterization of defective F1Fo ATP synthases, and ease of large-scale purification
procedures due to a practically unlimited supply of bacteria.
F1Fo ATP Synthase Mechanism
The overall function of F1Fo ATP synthase can be divided into three distinct parts:
proton conduction, coupling, and catalysis. The a and c subunits of Fo are responsible for
the translocation of protons through the membrane. The Fl y and a subunits of the rotor
stalk are responsible for coupling the energy acquired from the proton gradient to the Fl
catalytic sites. And the three catalytic sites located at the interfaces of the three oc and P
subunits are responsible for the synthesis of ATP or sometimes, as in the case of bacteria,
ATP hydrolysis. All three functions must be tightly integrated for the production of ATP.
Proton Translocation: Driving Rotation
The demonstration that the electrochemical gradient of protons drives the rotation
of bacterial flagella (245) in combination with Peter Mitchell's chemiosmotic theory (2)
began the search for evidence of rotation in F1Fo ATP synthesis. At the same time, a
model for proton transport was suggested by Cox et al. (212) and Boyer developed his
ideas for the binding change mechanism (discussed below) (246). But an indication of
rotational catalysis was not evident until the high-resolution crystal structure of F1
became available (20). This was followed a few years later by the first direct observation
of rotation when Noji et al. fixed the top of F1 to a glass coverslip and attached a
fluorescently labeled actin filament to the y subunit (15). Upon addition of ATP, rotation
of the actin filament was observed under an optical microscope at 0.2-10 revolutions per
second. At low concentrations of ATP (<600 nM), the actin filaments were observed to
rotate in a step-wise manner at 1200 intervals, which reflects the three catalytic sites in
the F1 003 3 hexamer (247). Experiments, in which two phenylalanine residues in the
nucleotide binding pocket were mutated to reduce the binding affinity of ATP, indicated
that the binding and hydrolysis of ATP is initially accompanied by a 900 substep
followed by a 300 substep attributed to product release (248, 249). These observations
were consistent with the two proposed "catches" observed between the P and y subunits
in the high resolution structure (see "The y subunit:" above) (20). The observance of the
rotation of the s and c subunits at the same speed and direction, indicating that these three
subunits rotate in synchrony, forming the central rotary machinery of the enzyme
complex (65-67). The concept of rotational catalysis with the rotation of the Yscio
subunits relative to the OC3 3 hexamer is now well accepted.
Several models have been proposed for proton translocation. One of the earliest
models suggested a series of side chains spanning the lipid bilayer formed a "proton
wire", involving amino acid residues casp61, aarg210, aglu219 and ahis245, 18 Which the protons
"hop" from one side chain to another until it passes through the membrane (212, 226,
250). Other models include a water-filled "proton channel" model, formed by the
charged residues of the a and c subunits, in which hydronium ions (H30 ) pass through
the lipid bilayer (251) and a "proton carrier" model, in which a proton binds on the
exterior of the membrane followed by a conformational change that brings the complex
through the membrane and releases the proton on the outer surface (252, 253).
The current prevailing model suggests that protons, in the form of H30 enter a
half channel created by the a subunit (Figure 1-9) (254, 255). The H30' is then believed
to protonate one of the clo subunits at residue casp61, which is positioned near the center of
the lipid bilayer (131). Besides forming the proposed half channel, the a subunit is
thought to play another crucial role in the protonation of the casp61. The pKa of the casp61
carboxyl side chain is uncommonly high, which is likely due to its lipid environment
(125). The essential aarg210 TOSidue is thought to facilitate a pKa shift of the casp61
carboxyl side chain to a lower pKa form during proton translocation (14, 132). Either the
protonation of the carboxylate, or possibly the release of a proton from a previously
protonated casp61 into an exit channel housed by the a subunit, somehow promotes the
generation of torque (Figure 1-9). The torque produced by proton translocation is
believed to drive the rotation of the ring of clo subunits relative to the a and b subunits
during rotational catalysis. Translocation of three to four protons generates enough
torque energy to rotate the clo ring by 1200 and results in the synthesis of a single
molecule of ATP.
Figure 1-9. Model of proton translocation and torque generation in Fo. Subunits included
are y (green), E (orange), a (yellow), b2 (purple), clo (blue). Protons (red) are
traveling in the direction of ATP synthesis. Protons are believed to travel
through a half channel housed in the a subunit as hydronium, H30 An
essential residue located near the center of the lipid bilayer in the c subunit,
casp61, iS thought to be protonated as another proton exits through another
"exit" half channel located on the cytoplasmic side of the membrane. The
protonation/deprotonation drives the rotation of the clo ring. The model was
drawn from schemes proposed by Junge (254) and Vik and Antonio (255).
In F1Fo ATP synthase, the mechanism of energy coupling requires both the rotor
stalk and the peripheral stalk. Rotation of the ring of clo subunits consequently results in
the rotation of the entire rotor stalk, which is essential in coupling the energy obtained
from proton translocation to the synthesis of ATP in the catalytic sites of the u3 3
hexamer located over 100 A+ away. The role of the peripheral stalk is to hold the hexamer
in place while the amino- and carboxyl-terminal a-helices of the y subunit rotates within.
In a fully functional and coupled enzyme, the kinetics of proton translocation and ATP
synthesis are linked so that one cannot proceed without the other, and vise versa (256).
Mutational studies suggested that the polar loop of the clo subunits was involved in
the coupling function. In F1Fo ATP synthases incorporated with the cgln42 glu Subunit
mutant, Fl was found to bind normally to the Fo mutant, but the passive leakage of
protons through this complex was not prevented as in must be in coupled enzyme
complexes (257). Also, ATP was hydrolyzed normally by this mutant, but hydrolysis
was not coupled to active proton translocation. A similar uncoupled phenotype was
found in complexes incorporated with a carg41 lys mutant (258). Another mutation at the
same site, carg41 his, WaS found to prevent the binding of F1 to Fo. Thus, this loop region
in the c subunit appears to play essential roles in both the binding of F1 and the coupling
of proton translocation to ATP synthesis. Second site suppressor mutations to the
cgln42 glu mutation were found in the a subunit, specifically, Sglu31 gly,val,1ys, that recouped
proton translocation and ATP synthesis (98). Crosslinking studies of cysteine double
mutants found cross-linked products between Scys31 and ccys40, Ccys42 and ccys43 (99).
Moreover, a functional contribution of the y subunit in energy coupling was demonstrated
by y subunit mutants that uncoupled proton transport and ATP hydrolysis (259). Second
site suppressor mutations were found in other regions of the y subunit (260, 261).
Cysteine substitutions formed crosslinks between the y subunit, cys205o, and ccys40, Ccys42
and ccys43 (118). Crosslinking studies also provided evidence for an interaction between y
and p (75, 77). In addition, the a subunit has been cross-linked to both the Fl and Fo
subunits, via introduced cysteines, indicating that it spans the entire length of the stalk
with the y subunit (60, 99). Combined, these observations had suggested that the
coupling mechanism occurred by direct interactions of the c subunit loop regions and the
y and a rotor stalk of F1, which convey the proton gradient energy to the catalytic sites,
probably by direct interactions.
The crystal structure later confirmed these interactions (20). The crystal structure
displayed a strikingly asymmetrical Fl due to differences in the domains of the a and P
subunits and the interactions formed with the single y subunit (20). The obvious
asymmetric positioning of the coiled coil of the y subunit is a key feature to the
mechanics of the binding change mechanism (discussed below) of F1Fo ATP synthase.
Its large carboxyl terminus a-helix passes through a hydrophobic sleeve formed by six
proline-rich loops of the a and P subunits, presumably resulting in the conformational
changes occurring in the catalytic sites (20). In the PE Subunit (see above), several
hydrogen bonds are formed with the y subunit which forms a "catch", resulting in
conformational changes. Specifically, residues yarg254 and ygln255 in the carboxyl terminal
helix form hydrogen bongs with PE-asp317, PE-thr318 and PE-asp319. Also, a second "catch" is
formed between the carboxyl terminal domain of the PT Subunit and the short helix of the
y subunit. Hydrogen bonds form between ylsss, Yys,90 and Yalaso within the PDELSEED
region, PT-asp394 and PT-glu398. Structural information suggests the two antiparallel coiled
coil a-helices of the y subunit may unwind during rotational catalysis and the a subunit
rotates around the Fl axis while undertaking a net translation of about 23 A+ (85). It is
likely that these gross changes observed in the structures revealed individual functional
states of the enzyme complex during catalysis.
Catalysis: The Binding Change Mechanism
F1Fo ATP synthases house three catalytic sites, located at the u3 3 interfaces, with
the predominate sites positioned in the P subunit and some contributions made by the a
subunits (20). The minimal complex capable of normal ATP hydrolysis activity is the
a3 37 complex (262). Boyer predicted that ATP synthesis requires a chronological
involvement of the three catalytic sites, each of which changes its binding affinity for the
substrates and products as it continues through a cyclical mechanism, referred to as the
"binding change mechanism" (246, 263). The mechanism of ATP synthesis, in terms of
the u3, P3 and Y subunits and the substrates, ADP and Pi, is described here (Figure 1-10).
The principles of the binding change mechanism have become the most commonly
used model for recounting the means of ATP synthesis by F1Fo ATP synthase. The three
distinct catalytic sites were described as the tight (T) site containing ATP, an empty open
(0) site, and the loose (L) site illustrated with ADP and Pi bound (Figure 1-10).
According to Boyer, catalysis starts with the binding of ADP and Pi at the open site. The
energy input from proton translocation drives the rotation of the rotor stalk, which
ultimately results in the conformational changes responsible for ATP synthesis such that
the tight site is converted to an open site, the open site assumes a loose site conformation,
and the loose sites becomes a tight sight. ATP is formed in the new tight site and the
molecule of ATP that was found in the original tight site is released and the binding
change mechanism starts fresh (Figure 1-10). Boyer's mechanism included three
proposals: i) only one site is actively synthesizing ATP at any given moment, ii) the
reaction occurs reversibly at this site, and iii) energy input is required to bind the
substrates, ADP and Pi, into the catalytic sites and to release the synthesized ATP, but not
for the actual reaction to occur. Strong evidence for this model has come from the crystal
structure of Fl (20) and the observance of rotation via an attached fluorescent actin
filament (discussed previously) (15, 65-67).
ATP Hydrolys s
Figure 1-10. The binding change mechanism. This is a simplified model of a more
detailed enzymatic mechanism described by Weber and Senior (189) in which
all three catalytic sites are transiently filled with nucleotide during ATP
synthesis. Subunits included are oc (red), P (green) and y (pink).
Genetic Expression and Assembly
The E. coli F1Fo ATP synthase is encoded in the 7 kb unc operon, which was
cloned and sequenced in its entirety (224, 264). The genes encoded are called the uncB,
unZcE, uncF, uncH, uncA, uncG, uncD and uncC coding for the a, c, b, 6, a, y, P and a
subunits, respectively. A single copy of each gene is transcribed into a single
polycysternic mRNA transcript, from which multiple polypeptides can be translated
(265-267). The synthesis of the correct number of subunits, resulting in the
stoichiometry of u3 38ysab2C10, iS thought to occur by translational regulation (264, 268).
The efficiency at which the individual subunits are synthesized were observed to be
variable and roughly corresponded to the stoichiometry of the intact enzyme complex
Far less is understood about the assembly of the complex F1Fo ATP synthase
enzyme compared to the structural and mechanistic studies. Some have proposed that no
particular pathway is necessary based on the observation that the complex can be
dissembled into individual subunits and then reconstituted in vitro (113, 148, 271). On
the contrary, some believe that the assembly follows an integrated pathway in vivo (272).
The idea of an assembly pathway appealed to many since it would prevent a newly
assembled Fl sectors from freely hydrolyzing cellular ATP in the cytoplasm, and newly
assembled Fo sectors from acting as open proton pores in the membrane. If assembly
were a random event, the potential to create isolated Fl and Fo sectors would exist. Some
evidence supporting the integrated pathway does exist. The a subunit is thought to
function as an inhibitor of ATP hydrolysis activity, undertaking large conformational
changes to allow the enzyme to switch from ATP synthesis to ATP hydrolysis under
conditions of low ATP or low proton gradient, respectively (102, 177, 273-276). Its
inhibitory action may possibly act to inhibit free ATPase activity in the partially
assembled state. Some have speculated that the binding of the Fl subunits to Fo may
influence the opening of the Fo proton channel in vivo (176, 277-279). The a subunit was
observed to be absent from membranes of cells lacking the b or c subunits (150).
Furthermore, work accomplished in our laboratory has showed that the b subunit
monomer does not integrate into the membrane has no affinity for Fl, indicating that the
formation of the b dimer in the membrane is an early event in F1Fo ATP synthase
assembly (199, 203, 218).
There is now ample evidence indicating that F1Fo ATP synthases are composed of
three functional parts, the catalytic core, the rotor stalk and the stator stalk. In the E. coli
enzyme complex, the catalytic core consists of the u3 3y Subunits and functions as an
ATP synthase or an ATPase. The rotor stalk consists of the ysclo subunits and couples
the energy of proton conduction to the synthesis of ATP by rotating within the u3 3
hexamer. Finally, the stator stalk consists of the Sab2 Subunits and remains in a fixed
position, anchored to the membrane by the a and b2 Subunits and to the u3 3 hexamer via
interactions made by the 6 subunit. Much of what was known of F1Fo ATP synthase has
been irrefutably confirmed by high resolution structures of partial complexes or model
polypeptides. To date, a high resolution structure of the complete enzyme, or at least the
complete Fo is eagerly anticipated.
Since the visualization of the peripheral stalk less than a decade ago a plethora of
data characterizing the b2 homodimer has emerged. The observation that the b2
homodimer was likely not a rigid structure, and possibly more of an elastic structure,
created the foundation for the work described in the following chapters. The work
illustrated in Chapters 2, 3, 4, 5 and 6 of this dissertation will characterize the role of the
peripheral stalk' s dimer of identical b subunits in the E. coli F1Fo ATP synthase by using
a combination of site-directed mutagenesis and biochemical methods. Chapter 2
demonstrates the ability of the E. coli b subunit to form heterodimers and the capability
of F1Fo ATP synthase complex to tolerate the incorporation of two different length b
subunits with a size difference of at least 14 amino acids (195). Chapter 3 demonstrates
the formation of b heterodimers including at least one and up to two defective b subunits
and documents indisputable evidence that F1Fo ATP synthases incorporated with b
subunit heterodimers are functional (280). Furthermore, the work accomplished in the
chapter indicates, for the first time, that each of the two b subunits makes a unique
contribution to the functions of the peripheral stalk, such that one mutant b subunit is
making up for what the other is lacking. Chapter 4 describes cysteine chemical
modifications constructed in the 6 subunit and shortened, lengthened and wild-type
length b subunits. The unc operon expression plasmids generated in this study will be
used in future fluorescent labeling experiments. Chapter 5 documents mutagenesis
experiments conducted on the extreme amino- and carboxyl termini of the b subunit
(202) (Bhatt et al., manuscript in preparation, 2004). Finally, Chapter 6 summarizes the
conclusions of this study and suggests the future directions that the work described in this
dissertation has offered.
INTEGRATION OF UNEQUAL LENGTH b SUBUNITS INTO F1Fo ATP SYNTHASE
F1Fo ATP synthases provide the bulk of cellular energy production in both
eukaryotes and prokaryotes (3, 5, 6). Enzymes in this family utilize the electrochemical
gradient of protons across membranes in order to synthesize ATP from ADP and
inorganic phosphate in a coupled reaction (16). In Escherichia coli, F,Fo ATP synthase is
a complex enzyme composed of approximately twenty-two polypeptides with the
stoichiometry of u3 3y~sab2C10 (6, 7). The F, portion is composed of the subunits
a3 3yis and is responsible for enzymatic catalysis. The Fo portion of the enzyme consists
of the ab2C10 Subunits and is responsible for the translocation of protons through the
Electron microscopy has shown that the Fl and Fo sectors are linked by two slender
stalk structures (11). During ATP synthesis proton translocation drives the rotation of the
central stalk, which consists of subunits ys, within the u3 3 hexamer held stationary by
the peripheral stalk. This rotation propagates the conformational changes in the active
sites located at the up interfaces driving catalytic activity (3, 15, 62, 65, 281, 282). The 6
subunit of Fl and a dimer of two identical b subunits from Fo comprise the peripheral
stalk acting as the stator. The 6 subunit has been visualized seated at the top of the Fl
a3 3 hexamer (187). However, recent evidence has suggested that the 6 subunit may be
positioned slightly to the side of F1 in association with a single a subunit (100, 141-143).
The C-terminal region of the 6 subunit is in direct contact with the extreme C-terminal
end of the b dimer (3, 185, 200, 219, 220). The b subunit dimer constitutes the majority
of the peripheral stalk stretching from within the membrane to near the top of F1 (12).
Dimerization of the b subunits is required for the normal assembly and function of
F1Fo ATP synthase (199). The two b subunits are believed to exist in parallel as an
extended structure spanning from the periplasmic side of the membrane to near the top of
F1. Each has a N-terminal transmembrane domain, a tether domain extending from the
surface of the membrane to the bottom of F1, a dimerization domain and a 6-binding
domain (13). The ability of b to bind to F1 was proportional to the ability of b to form
dimers, suggesting the necessity of the b dimer formation before the binding of F1 to the
complex (199). Presently, there is no high-resolution structure of the entire b subunit. A
model polypeptide of the first 34 residues of the N-terminus has been solved by NMR,
revealing a hydrophobic membrane-spanning a-helix (136). A crystal structure of a
monomeric dimerization domain, consisting of residues 62-122, has been solved and
refined to 1.55 A (138). Dunn and coworkers have constructed a model in which the two
a-helices of the b62-122 TegiOn form a right-handed coiled coil. Much of the structural
information on the b dimer has been gleaned from classical biochemical approaches such
as CD-spectroscopy, crosslinking and sedimentation experiments (30, 166, 192, 196, 197,
203, 283). These studies revealed that the overall structure of the b subunit dimer is a
highly extended conformation with approximately 80% a-helix.
Previous studies have shown that b subunits with deletions of up to eleven amino
acids and insertions of up to fourteen amino acids, corresponding to approximately 16 A~
and 21 A+, respectively, formed functional FiFo complexes (193, 194). When b subunits
with either a seven amino acid deletion or an insertion, b or b 7, respectively, were
incorporated into the F Fo ATP synthase complex, the properties of the enzymes were
essentially wild type. These observations suggested that the role of the b dimer is more
of a flexible structural feature. However, it was not known whether this flexibility
extended to the dimerization of two b subunits of unequal lengths and their incorporation
into an enzyme complex. In the present study an experimental system was developed to
allow expression of two different b subunit genes and determine whether the differing b
subunits were assembled into an F1Fo ATP synthase complex. Here, we demonstrate that
the F1Fo ATP synthase complex can tolerate b subunits with a size difference of at least
14 amino acids.
Materials and Methods
Molecular biology enzymes and mutagenic oligonucleotides were obtained from
Invitrogen (Carlsbad, CA), Life Technologies, Inc. (Grand Island, NY), New England
Biolabs (Beverly, MA) and Stratagene (La Jolla, CA). Reagents were obtained from
Sigma (St. Louis, MO), BioRad Laboratories (Hercules, CA) and Fisher Scientific
(Pittsburgh, PA). Plasmid purification kits were acquired from Qiagen Inc. (Valencia,
CA). The anti-rabbit immunoglobulin horseradish peroxidase-linked whole antibody
(from donkey), anti-mouse immunoglobulin horseradish peroxidase-linked whole
antibody (from sheep), Hybond ECL Nitrocellulose membrane,
electrochemiluminescence Western blotting reagents and high performance
chemiluminescence film were purchased from Amersham Biosciences (Piscataway, NJ).
Polyclonal antibodies against SDS-denatured b subunit (284, 285) were generously
provided by Dr. Karlheinz Altendorf (Universitait Osnabriick, Osnabriick, Germany).
Mouse monoclonal antibodies against the peptide epitope of hemagglutinin protein of
human influenza virus (HA epitope tag) were purchased from Roche Molecular
Biochemicals (Indianapolis, IN). Monoclonal antibodies against the epitope found in the
P and V proteins of the paramyxovirus, SV5 (V5 epitope tag) were purchased from
Strains and Media
The bacterial strains used to create the epitope tagged b subunits include the wild
type b subunit expression plasmid, pKAM14, and plasmids used to express b subunits
shortened or lengthened by 7 amino acids, pAUL3 and pAUL19, respectively, and have
been described previously (193, 194, 203). The plasmids encoding the different uncF(b)
genes were used to compliment E. coli strain KM2 (Ab) carrying a chromosomal deletion
of the gene (218). All strains were streaked onto plates containing Minimal A media
supplemented with succinate (0.2% w/v), to determine enzyme viability. Cells harvested
for membrane preparation were grown in Luria Bertani media supplemented with glucose
(0.2% w/v) (LBG). Isopropyl-1 -thio-P-D-galactopyranoside (IPTG)(40 Clg/ml),
ampicillin (Ap) (100Clg/ml), and chloramphenicol (Cm) (25 Clg/ml) were added to media
as needed. All cultures were incubated at 37oC for the appropriate duration.
Recombinant DNA Techniques
Plasmid purification. Plasmid DNA was purified with the Qiagen Mini-Prep and
Maxi-Prep kits according to the protocols provided from the manufacturer. Mini-preps
required 4 mL (high copy plasmid) or 6 mL (low copy plasmid) of an overnight bacterial
culture carrying the desired plasmid. Maxi-preps required a 500 mL culture grown
overnight. Final elution volumes for the mini-prep kit was 30 or 50 C1L, for low copy and
high copy plasmids, respectively, and 200-250 C1L for the maxi-prep kit. Final
concentrations of 0.5 and 1.0 Clg/CIl plasmid were routinely obtained with the Qiagen mini
and maxi-prep kits.
Digestions, ligations, and transformations. Restriction endonuclease digestions,
ligations, and transformations were performed according to the recommendations of the
manufacturers (New England Biolabs, Stratagene and Invitrogen). For analytical
purposes, restriction endonuclease digestions were normally prepared in a total volume of
20 CIL, including plasmid DNA, enzyme, buffer, ddH20 and occasionally BSA, and then
incubated for an hour at the temperature specified by the manufacturer. Ligations
required two purified double-stranded DNA fragments, a vector and an insert, of known
length and concentration (ng/C1L). DNA fragments were routinely separated in a 0.8 %
agarose gel by electrophoresis and the appropriate sized fragment was excised and
purified using a Qiagen, Inc. QIAquick Gel Extraction kit. The vector fragment
contained the antibiotic resistance gene and the origin of replication. The insert typically
contained the desired gene or a site specific mutation. Two control reactions and two
ligation reactions were set up. The typical reaction was set up in 20-40 CIL and included
vector, insert, ATP, T4 DNA ligase buffer, T4 DNA ligase and ddH20. The first control
reaction was a control for uncut plasmids, containing no insert and no ligase, and the
second, containing no insert, was a control for the vector' s ability to ligate with itself.
The femptomolar concentration (fmol/CIL) was determined from the known size and
measured concentration. Two ligation reactions were then set up, the first had 1 part
vector to 3 parts insert and the second was 1:10. Typically 5-15 fmol of vector was used.
The four reactions were incubated for 5 minutes at room temperature if T4 High
Concentration DNA ligase was used or overnight at 16 oC if T4 DNA ligase was used and
then transformed into competent E. coli. Transformations were performed in one of three
different E. coli strains. DH~a competent cells and XL10-Gold Ultracompetent cells
were competent bacteria purchased from Invitrogen and Stratagene. These bacteria were
used for purposes of plasmid preparations or to transform with a mutagenesis reaction
such as Quikchange or ligations. The DH~a and XL10-Gold bacteria were stored at -80
oC. Basically, 25 CIL pre-aliquoted cells were thawed on ice and 1-5 CIL plasmid DNA
was added and mixed by gentle stirring with the pipette tip. The bacteria were incubated
on ice for 30 minutes, heat shocked at 42 oC for 45 seconds, and then incubated on ice for
2 minutes. 1 mL LBG was added and then incubated at 37 oC taped to a roller drum. The
cells were harvested by centrifugation for 1 minute, the supernatant was discarded, the
bacteria were resuspended in the remaining media, spread onto a LBG plate
supplemented with the appropriate antibiotic and incubated at 37 oC overnight. XL10-
Gold cells required treatment with P-mercaptoethanol (P-ME) before transformation to
increase the efficiency. 2 C1L of the provided P-ME mix was added to 45 CIL pre-
aliquoted cells and incubated on ice for 10 minutes prior to transformation with gentle
swirling every 2 minutes. The transformation proceeded as described above. KM2 (Ab)
was a strain of E coli generated by a previous lab member (218) and maintained in our
laboratory. This strain was used for purposes of plasmid expression and the study of the
b subunit of F1Fo ATP synthase. Before transformation, KM2 had to be made competent.
Sterile technique was important since KM2 cells cannot be selected for by an antibiotic.
All reagents and supplies were pre-chilled unless otherwise noted. A culture of KM2
cells was grown overnight at 37 oC in 5 mL LBG. The overnight culture (100 CIL) was
inoculated into 10 mL pre-warmed (37 oC) LBG and allowed to grow for 2-4 hours. The
fresh culture was poured into sterile polypropylene tubes and incubated on ice for 10
minutes. The bacteria were then harvested by centrifugation (6,000 rpm in a ss-34 rotor)
for 10 minutes. The supernatant was discarded and the tubes were inverted on a
Kimwipe for 1 minute to allow excess LBG to drain. The bacteria pellet was
resuspended in 10 mL cold, sterile 50 mM CaCl2 and incubated on ice for 45-60 minutes.
The bacteria were harvested as described above, resuspended in 1 mL of the 50 mM
CaCl2, and stored at 4 oC until use. Generally, the competent KM2 cells could be used
after two hours on ice but were most efficient for transformation at 24 hours and expired
at 72 hours. For transformation, 100-200 CIL of cells were used as previously described.
Site-directed mutagenesis. Site-directed mutagenesis was performed either by
means of a Stratagene Quikchange XL kit or by ligation-mediated mutagenesis.
Oligonucleotides containing the desired mutation(s) were designed to anneal to the same
sequence on opposite strands of the plasmid (sense and antisense primers) (Appendix A)
(Figure 2-1). When possible, a silent mutation was encoded to add or delete a restriction
endonuclease recognition sequence to allow for easy screening of the mutation. Primers
were optimally designed by ensuring the mutation was in the middle of the sequence, a
cytosine (C) or guanine (G) flanked both ends of the sequence, and the melting
temperature (Tm) was greater than or equal to 78 oC. The Tm was calculated as
Tm=81.5+0.41 (%GC)-675/N-%mismatch, where N was the primer length (bases). When
introducing insertions or deletions, "%mismatch" was dropped from the formula.
Mutagenic Primers (sense strand)
histidine tag Sphl H H H
HA tag Y P Y D V
5' CTAAATAGAGGCATTGTGCTATGTACCCATATGAC GTG
P D Y A
C CGGAC TACGCGAcATC-TTAcACGCAAICAATCCTCGGC 3'
V5 tag SacI G K P I P N P
5' CGTGGATAAAC-TTGTCGCTGAGCTC GGTAAAC CGATCC CGAAC CCG
L L G L D S
CTGCTGGGTCTGGACTCTAC CTAAGGAGGGAGGGGCTGATGTCT 3'
+ Sphl site Sphl
5' CTGTAAGGAGGGAGGGGCATGC GTCTGAATTTTA-TTACG 3'
-Sphl site +ATG H H H
5' GGAGGGAGG GGCTGATGCACCACCAC 3'
Figure 2-1. Oligonucleotides for epitope tags and mutagenesis of uncF(b). Shown are the
sense strands of the mutagenic primer pairs. Green and red codons specify
translation start and stop sites, respectively. A) Mutagenic primers encoding
the sequence of the desired epitope tag insertion along with a silent mutation
that introduced a new endonuclease recognition sequence. Epitope tags were
inserted as described in the Materials and Methods. The oligonucleotides
sequences specifying the histidine, HA or V5 epitope (bold blue) tag are
labeled with the corresponding amino acid. The Sphl, Ndel and SacI
restriction sites (underlined) were added along with the histidine, HA and V5
tags, respectively, to facilitate screening. B) Oligonucleotides designed to
construct the one-plasmid expression system.
Primers were not fast polynucleotide liquid chromatography (FPLC) or polyacrylamide
gel electrophoresis (PAGE) purified as called for in the protocol. The reaction mixture
consisted of 5 CIL Stratagene's 10X Pfu buffer, 3 CIL QuikSolution, 50 ng wild type
plasmid, 125 ng each of the sense and antisense oligonucleotides, 1 C1L of 25 mM
dNTP's, and 1 unit Pfu turbo polymerase. The total volume was brought up to 50 C1L
with ddH20 and PCR was performed in a Perkin Elmer GeneAmp PCR System 2400
thermocycler according to the following cycling parameters: 950C 1 minute presoak; 18
cycles of 950C 50 seconds, 600C 50 seconds, 680C 1 minute per kb of plasmid length;
40C 7 min. Upon completion of the thermocycling, 1 C1L DpnI restriction endonuclease
was added to the reactions in order to digest the methylated (nonmutated) plasmid DNA.
The plasmids carrying the desired mutation were then transformed into competent DH500
cells, purchased from Life Technologies, and grown on LBG plates supplemented with
the appropriate antibiotic. Plasmids carrying the desired mutation(s) were screened for
by restriction endonuclease analysis and then the nucleotide sequences were directly
determined by automated sequencing in the core facility of the University of Florida
Interdisciplinary Center for Biotechnology Research (ICBR).
Mutagenesis and Strain Construction
Plasmids pKAM14 (b, Apr) (203), pAUL3 (ba7, Apr) (194) or pAUL19 (b 7, Apr)
(193) were used to construct the epitope tagged b subunits. Epitope tags were inserted
into each of the plasmids using the Stratagene Quikchange kit. A histidine epitope tag
was inserted at the N-terminus by mutagenesis between the first and second codons of the
uncF(b) gene to express bhis or b+7-his (Appendix A) (Figure 2-1). Plasmids pTAM37
(bhis) and pTAM3 5 (b+7-his) were created by digesting both of the recombinant histidine-
tagged b subunit plasmids with PstI and Ndel and subsequently lighting the genes into a
plasmid conferring the chloramphenicol resistance gene and the pACYC 184 origin of
replication (Table 2-1). Likewise, an HA epitope tag was inserted at the N-terminus by
mutagenesis between the first and second codons of the uncF(b) gene to generate
plasmids pTAM36 (bHA) Or pTAM34 (bA7-HA). A V5 epitope tag was added to the C-
terminus by site-directed mutagenesis before the termination codon of the uncF(b) gene
to express bvS or ba7~vs from plasmids pTAM46 and pTAM47 (Figure 2-1). The
recombinant HA-tagged and V5-tagged b subunit plasmids included the ampicillin
resistance gene and the pUC18 origin of replication. Unique restriction enzyme sites
Sphl, Ndel and SacI were constructed near the histidine, HA and V5 epitope tag
sequence, respectively, for an initial detection of the insertions, and then the nucleotide
sequence was subsequently confirmed by automated sequencing in the ICBR core
facility. Additionally, a set of plasmids was designed in order to express two different-
tagged b subunits from a single transcript (Figure 2-2). As an example, a unique
restriction enzyme site, Sphl, was created in conjunction with the histidine tag between
the Shine Dalgarno sequence and the first codon of uncF(b) to create pTAM3 5 (Figure 2-
2A). In a separate site-directed mutagenesis reaction, an Sphl site was added to the
pTAM34 plasmid downstream of the HA-tagged b subunit and behind another favorable
Shine Dalgarno sequence (Figure 2-2B). The two plasmids were digested with Sphl and
BstEII restriction endonucleases. The 3.2 kb vector fragment from pTAM34 and the 623
bp insert fragment from pTAM3 5 were isolated and then excised from a 0.8 % (w/v)
agarose gel and purified with a QIAquick gel extraction kit. The vector and insert
fragments were then allowed to ligate overnight at 16 oC. It was then crucial to mutate an
A ori B
Cm' L 4
Sph| HA tag
6.9 kb I~H) SD~td~h
(b+7-hi SD Ofl (b-y+AdSh
3 1 kb Ligate 623 bp
C HA tag
rp SD b,
3.7 kb SD
(bH -~ _HA & +7-his
Figure 2-2. Construction of the single transcript expression system. A) A unique
restriction enzyme site, Sphl, was created in conjunction with the histidine tag
between the Shine Dalgarno sequence and the first codon of uncF(b) to create
pTAM3. B) An Sphl site was added to pTAM34 downstream of the HA-
tagged b subunit and behind another favorable Shine Dalgarno sequence. The
two plasmids were digested with Sphl and BstEII. The 3.2 kb vector fragment
from pTAM34 and the 623 bp insert fragment from pTAM3 5 were ligated.
Additional mutagenesis (see Materials and Methods) resulted in C) a 3.7 kb
plasmid, pTAM40, which expressed bA7-HA and b+7-his from the same promoter
and included the ampicillin resistance gene and the pUC 18 origin of
replication. In similar constructions, plasmids pTAM41 (bwt-HA and b+7-his)
and pTAM42 (bwt-HA and bwt-his) were created.
intrinsic added start codon, which is in the Sphl recognition sequence, to prevent a
missense mutation. The mutagenic oligonucleotide was designed to accomplish three
tasks in one reaction: 1) mutate the "ATG" found in the Sphl recognition sequence, 2)
add a new Shine Dalgarno sequence in a favorable position from the true start codon, and
3) mutate the "GTG" start codon to a more favorable "ATG" start site (Figure 2-1). The
resulting 3.7 kb plasmid, pTAM40, expressed bA7-HA and b+7-his from the same promoter
and included the ampicillin resistance gene and the pUC 18 origin of replication (Figure
2-2C). In similar constructions, plasmids pTAM41 (bwt-HA and b+7-his) and pTAM42 (bwt-
HA and bwt-his) were created (Table 2-1). Throughout this dissertation, the insertion or
deletion and the epitope tag are indicated after the plasmid name for clarity, for example
plasmid pTAM35 (b+7-his). Each plasmid and the control plasmids pKAM14 (b) and
pBR322 were expressed in the E. coli cell line KM2 (Ab) for study, so that the only b
subunits in the cells were the product of the plasmid genes.
The two plasmid expression system successfully allowed expression of various
combinations of histidine tagged and V5-tagged b subunits in the same cell (Figure 2-7).
Appropriate antibiotics were added to the growth medium, and in the case of the
coexpressed plasmids, both ampicillin and chloramphenicol were added to select for cells
expressing both plasmids.
Crude Preparative Procedures
Inverted membrane vesicles from KM2(Ab) strains expressing the desired b subunit
epitope tagged F1Fo ATP synthase complex were prepared for activity assays, Ni-resin
purification and Western Blot analysis. Unless otherwise noted, all reagents, rotors and
materials were kept at 4 oC. The membrane preparations were prepared by inoculating a
10 mL starter culture, grown overnight, into a 2 L Erlenmeyer flask containing 500 mL
LBG, supplemented with the appropriate antibiotic, ampicillin (Ap) and/or
chloramphenicol (Cm). Similarly, 1 mL of a starter culture was inoculated into a nephalo
flask containing 50 mL LBG (Ap and/or Cm). The bacteria were grown at 37 oC in a
New Brunswick Scientific incubator shaker (220 rpm) and the turbidity was monitored
using a Klett-Summerson photoelectric colorimeter. IPTG (40 CIM) was added when the
turbidity reached 75 Klett units and the cells were collected when the turbidity reached
150 Klett units. The bacteria were harvested by centrifugation for 10 minutes at 8,000Xg
in a Sorvall GSA rotor. The pellets were rinsed once with TM buffer (50 mM tris-HC1,
10 mM MgSO4, pH 7.5) and then resuspended in a final volume of 10 mL TM buffer.
DNasel (10 mg/mL) was added to a final concentration of 10 Clg/mL and the bacteria
were broken by passing through a French Pressure Cell one time at 14,000 psi. Cellular
debris and unbroken cells were removed by centrifuging twice at 10,000Xg for 10
minutes. Membranes were then collected by ultracentrifugation at 150,000Xg in a
Beckman 70.1 Ti rotor for 1.5 hours. The membrane pellets were rinsed once with TM
buffer and then resuspended in TM buffer to a final volume of 2 mL using a 2 mL
Wheaton tissue grinder. For the purposes of Western blot analysis or Ni-resin
purification, one ultracentrifugation step sufficed. However, activity analysis required an
additional ultracentrifugation step in order to remove nonspecifically bound ATPases. In
this case, the membrane pellets were resuspended in a final volume of 10 mL using a 10
mL Wheaton tissue grinder and the ultracentrifugation step was duplicated.