The a subunit of Escherichia coli F1F0 ATP synthase as a model for investigating proton translocation


Material Information

The a subunit of Escherichia coli F1F0 ATP synthase as a model for investigating proton translocation
Physical Description:
x, 197 leaves : ill. ; 29 cm.
Hartzog, Phillip Edwin, 1964-
Publication Date:


Subjects / Keywords:
Proton-Translocating ATPases -- genetics   ( mesh )
Escherichia coli   ( mesh )
Protons -- metabolism   ( mesh )
Mutagenesis, Site-Directed   ( mesh )
Models, Genetic   ( mesh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1994.
Includes bibliographical references (leaves 180-196).
Statement of Responsibility:
by Phillip Edwin Hartzog.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 50683433
System ID:

Full Text








I would like to thank all the members of my supervisory

committee, Dr. Michael Kilberg, Dr. Charles Allen, Dr.

Sheldon Schuster, Dr. Paul Hargrave, and Dr. Lonnie Ingram,

for their help and advice, no matter how far my project

strayed from the original proposal. Additionally, I would

like to thank Dr. Donald Allison for his expert advice and

help with data interpretation. I would also like to thank

Dr. Richard Boyce, because he always knew how to fix any

problem I might have with the University.

I would like to express my gratitude to my mentor, Dr.

Brian Cain, who has taught me how to start becoming a

scientist. Hopefully, one day I will get the chance to pass

on some of the knowledge he has taught me to an unsuspecting

graduate student.

And last, but not least, I would like to thank all my

friends and family, here and back in North Carolina; without

their friendship and support I would never have made it

through this process. Additional thanks goes to my long

time friend Bob Whitener, for without his financial support

when I needed it most, I would not have made it to





LIST OF FIGURES ...... ... vi


Introduction ..... . ... 1
Similarity Among FiFo ATPases . 3
Binding Change Mechanism . .. 17
E. coli as a Model for Other ATP Synthases 20
E. coli FiFo ATP Synthase . 26
The a Subunit of E. coli FFo ATP Synthase .. 40

Molecular Techniques . 51
Membrane Preparations . .. .. 67
Assays . .... 71

Introduction . .. 78
Results . . 80
Discussion . . 92

THROUGH a. IN THE a SUBUNIT .... .. 94
Introduction . ... 94
Results . . .. 98
Discussion . .. .. 110

Introduction . ... 116
Results . .. 119
Discussion . . 134


Introduction . . .. 137
Results . . 138
Discussion . ... 156

Overview of the a Subunit Missense Mutations 160
Modeling Mitochondrial Mutations 169
Coupling Proton Transport to ATP Catalysis 172




Table 2-1. List of Strains and Plasmids 53

Table 3-1. Complementation Analysis .. 82

Table 3-2. Membrane Associated ATPase Activities 85

Table 4-1. Mutations Generated in the uncB (a) Gene 100

Table 4-2. Growth Properties of a Subunit Mutants 102

Table 4-3. Properties of a Subunit Mutations at ag2 111

Table 5-1. Properties of the a7R Mutation .. 123

Table 5-2. Properties of the aLo7p Mutation 128

Table 6-1. Mutations Generated in the uncB (a) Gene .140

Table 6-2. Growth Properties of Mutant Strains 141

Table 6-3. Biochemical Properties of Mutant Strains .145


Figure 1-2. Diagram of the E. coli FIFo ATP synthase. 18

Figure 1-2. Model of the binding change mechanism and

its relationship to energy supplied by proton

translocation. . ... 21

Figure 1-3. Alignment of a-like subunits. 41

Figure 1-4. Proposed topology model of the a subunit. 48

Figure 2-1. Plasmid pPH12 (a) that was used to

construct site-directed mutations in the uncB (a)

gene by cassette mutagenesis. .... 62

Figure 3-1. Southern analysis of strain PH105 (a,)

and its parent strain 1100. ... 84

Figure 3-2. Energization of membrane vesicles

determined by fluorescence quenching of ACMA. 87

Figure 3-3. Western analysis of strain RH305 (ass)

and strain PH105 (a). . ... 89

Figure 4-1. Alignment of amino acid sequence in the

last proposed membrane spanning segment. ... 95

Figure 4-2. Oligonucleotides used for cassette

mutagenesis and the sequence from plasmid pPH12

(a). . . 99

Figure 4-3. Energization of membrane vesicles

determined by fluorescence quenching of ACMA. .106

Figure. 4-4. Proton impermeability of stripped

membranes from a subunit mutant strains. 108

Figure 5-1. Alignment of a-like (ATPase 6-like)

subunits. . ... 120

Figure 5-2. Plasmid pBDC26 (a7.WE21j9,K) and the

oligonucleotide cassettes used to make the mutant

plasmid pUNCB4.70 (agmR) and plasmid pUNCB4.71

(aL 7,.p) . . 121

Figure 5-3. FFo ATP synthase mediated ATP-driven

vesicle acidification for a s . 124

Figure 5-4. NADH driven ACMA fluorescence quenching of

ak ., stripped vehicles . .. 126

Figure 5-5. FFo ATP synthase mediated ATP-driven

vesicle acidification for a~,ro. ......... 130

Figure 5-6. NADH driven ACMA fluorescence quenching of

ak- stripped vesicles . .. 131

Figure 5-7. Western Analysis of ao7_p comparing intact

verses stripped vesicles . ... .133

Figure 6-1. Alignment of the sequence variations seen

at aG218 aE219 and aH2 .... .. .. 139

Figure 6-2. Western analysis of mutant strains.

Stripped membranes prepared from mutant strains

were used for immunoblotting with anti-b

antibody. . ....... 147


Figure 6-3. ATP driven energization of membrane

vesicles determined by ACA. . .

Figure 6-4. NADH driven energization of stripped

membrane vesicles determined by ACMA .

Figure 6-5. ATP driven energization of membrane

vesicles under different pH conditions. .

Figure 6-6. ATP driven energization of membrane

vesicles under different ATP concentrations. .

Figure 7-1. Compilation of missense mutations in the a

subunit . . .







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



Phillip Edwin Hartzog

April 1994

Chairperson: Dr. Brian D. Cain
Major Department: Biochemistry and Molecular Biology

FFo adenosine triphosphate (ATP) synthases use an

electrochemical gradient of protons across energy

transducing membranes to synthesize ATP. The enzyme complex

consists of a Fi sector containing the catalytic activity,

and a Fo sector capable of proton conductance. All FIFo ATP

synthases contain an essential a-like subunit in the Fo

sector. Primary structural homology with other a-like

subunits indicates conservation of amino acid positions

ae2, aF56, a759, and an. These amino acid positions were

investigated using site-directed mutagenesis techniques, and

analysis of the mutant strains indicated these amino acids

have an important structural role in proton translocation.

The validity of comparing studies in E. coli to all

FFo ATP synthases was demonstrated by the modeling of the

mutation proposed to cause the human mitochondrial genetic

disease neurogenic muscle weakness, ataxia and retinitis

pigmentosa. The biochemical effects of two mutations

occurring in the human ATPase 6 subunit were determined by

constructing the comparable mutations in the a subunit of E.

coli. Both mutations greatly decreased FIFo ATP synthase

activity by destabilizing the enzyme complex.

E. coli was also used to model the a-like subunits from

chloroplast and alkaliphilic bacteria FIFo ATP synthases.

The a-like subunits from both types of ATP synthases contain

variations from the E. coli amino acid sequence at positions

aG18 and a,45. The double mutations aG2,& -Du and aG1nKmH24.G

introduce substitutions which mimic the amino acids present

in chloroplast and alkaliphilic bacteria, respectively. The

enzymes carrying the double substitutions were functional at

near wild-type levels, while those with the single

substitutions ams.G, aG2D and aG2-K had greatly reduced

enzyme function. The phenotypes exhibited by these mutants

also indicated an involvement in the coupling of proton

translocation to ATP synthesis.

The a subunit from E. coli was shown to be a good model

for studying changes in a-like subunits from different FIFo

ATP synthases. The study of these changes has led to a new

understanding of how the translocation of protons is coupled

to the synthesis of ATP.



Oxidative phosphorylation (OXPHOS) is the primary

bioenergetic pathway for adenosine triphosphate (ATP)

synthesis in aerobic organisms. The electron transport

chain generates an electrochemical gradient of protons which

is used to drive ATP production by FIFo ATP synthase. FIFo

ATP synthase harnesses the energy of the proton

electromotive force and uses it to synthesize ATP from ADP

and Pi. The coupling of an electrochemical gradient with

the synthesis of high energy chemical bonds connects the two

major forms of stored energy in living organisms. In this

capacity FIFo ATP synthase plays a central role in


The chemiosmotic hypothesis presented by Peter Mitchell

(Mitchell, 1961) proposed a mechanism of ATP production from

an electrochemical gradient of protons. FIFo ATP synthase

was clearly demonstrated to couple proton translocation to

ATP synthesis by Racker and Stoeckenius (1974). Intense

study of FIFo ATP synthase for more than four decades has

revealed a great deal of structural and functional

information about the enzyme. Current reviews on this topic

have been presented by Boyer (1987, 1993), Fillingame

(1990a, b), Futai et al. (1989), Hatefi and Matsuno-Yagi

(1992), Penefsky and Cross (1991), Schneider and Altendorf

(1987), and Senior (1990). The structural and functional

analysis remains incomplete. FIFo ATP synthase is composed

of two sectors, F1 and Fo, with defined functional

activities. F, is easily dissociated from the membrane and

retains the ATP hydrolysis activity. The mechanism of ATP

hydrolysis in the F, sector of the enzyme has been studied

in great detail (reviewed in Cross, 1992; Senior, 1990), and

considerable structural data exist for this sector (Abrahams

et al., 1993; Akey et al., 1983; Tsuprun et al., 1984;

Tiedge et al., 1985; Boekema et al., 1986, 1988; Gogol et

al., 1989a, b; Ishii et al., 1993). However, much less is

known about the mechanism of proton translocation through

the Fo sector, and how the energy of proton translocation is

coupled to ATP synthesis. The structural information for

the Fo sector is also limited. The F, and Fo sectors are

connected by a stalk that has been proposed to couple proton

translocation to ATP synthesis. This work focuses on the

mechanism of proton translocation, and indirectly, its

coupling to the catalytic activities.

The facultative bacterium E. coli offers many

advantages in the study of FFo ATP synthase. The ability


of E. coli to grow via glycolysis allows the construction of

strains defective in FIFo ATP synthase. Molecular

techniques can be employed to alter genes on the chromosome,

as well as, express FIFo ATP synthase subunits from

plasmids. E. coli is amendable to growth analysis under a

wide variety of conditions including different carbon

sources and temperatures, as well as both, anaerobic and

aerobic environments.

Similarity Among FiF, ATPases

Biological Similarities

FIFo ATP synthase is used by all aerobic organisms to

produce ATP from the electrochemical gradient of protons

produced by an electron transport chain (Redox) or

photoreaction center (Photo-redox). FIFo ATP synthases are

localized to the inner mitochondrial membranes of

eucaryotes, chloroplast thylakoid membranes of plants, and

the cytoplasmic membranes of procaryotic organisms. FIFo

ATP synthases from procaryotes to eucaryotes bear a general

structural resemblance and apparently have comparable

mechanisms for ATP catalysis and proton translocation

(Senior, 1990).

Vacuolar ATPases. Under anaerobic conditions

facultative organisms can use FIFo ATP synthase in reverse

to produce a proton gradient by hydrolysis of ATP. This

allows the maintenance of the membrane potential important

for many biological membrane functions. Similarly,

anaerobic organisms are believed to use an enzyme akin to

FFo ATP synthases for maintaining membrane potential.

Vacuolar ATPases (V-type ATPases) of eucaryotic organisms

are believed to be derived from anaerobic F-type ATPases,

and consequently share homology with FIFo ATP synthases at

the amino acid sequence, structural, and functional levels

(SUdhof et al., 1989; Stone et al., 1990; Senior, 1990).

Propionigenium modestum. An example of the distant

evolutionary relationship among the F-type ATPases is

demonstrated by the strict anaerobic bacterium

Propionigenium modestum which uses an enzyme related to FIFo

ATP synthase to produce ATP by the translocation of Na+ ions

(for review see Dimroth, 1987, 1992). Laubinger and Dimroth

(1989) found that P. modestum FIFo ATP synthase would pump

protons under conditions of low Na+ concentrations.

Laubinger et al. (1990) reported that E. coli Fi sector

would bind to P. modestum Fo sector and function in a

reconstituted system. Additionally Kaim et al. (1992)

reported the expression of P. modestum Fo subunits in a E.

coli strain deleted of Fo subunits, mimicking the earlier in

vitro experiment. This demonstrated that functional and

structural interactions in FIFo ATP synthases have been

conserved between widely divergent organisms. As the above

study suggests, FFo ATP synthase subunits of P. modestum

were found to exhibit evolutionary conservation at the

primary amino acid sequence level with other FIFo ATP

synthases (Amann et al., 1988; Ludwig et al., 1990; Esser et

al., 1990; Woese et al., 1990; Kaim et al., 1992).

The similarities between both E. coli and P. modestum

in coupling of ATP synthesis with ion translocation reveals

a relationship between the translocation of protons and the

translocation of other ions. The implication is that

studies of P. modestum enzyme are applicable to the E. coli

enzyme, as well as all other FFo ATP synthases. If the

mechanism of ion translocation is conserved between P.

modestum and E. coli, then it is reasonable to propose that

E. coli FIFo ATP synthase transports hydronium ions (H3O*)

instead of protons (H+). Hydronium ions are similar to

hydrated Na+ ions at the chemical level, and therefore will

exhibit similar binding characteristics (Boyer, 1988;

Laubinger et al., 1990). The ability of the P. modestum

FFo ATP synthase to transport protons as well as Na+ ions

is conducive to a hypothetical ion binding site that could

have a specificity for H3O+ or hydrated Na+ ions. This

proposed ion binding site of P. modestum FIFo ATP synthase

has a higher affinity for hydrated Na+ ions, but binds H3O+

ions in the absence of Na+ (<1 mM). However, the E. coli

FIFo ATP synthase ion binding site appears specific for H3O+

ions (> 106 affinity for H30+ ion) since it exhibits no

ability to transport Na+ (Laubinger & Dimroth, 1989). A

proposed mechanism of ion transport in FIFo ATP synthase is

discussed below.

Chloroplast and cyanobacteria. Another example of

structural and functional conservation was reported by Lill

et al. (1993). The genes encoding F, subunits from spinach

chloroplast and the cyanobacterium Synechocystis sp. PCC

6803 were expressed in E. coli strains that carried

mutations in the corresponding F, subunits. The strains

expressing chloroplast and cyanobacteria F, subunits were

then grown on a succinate minimal medium to determine if the

chloroplast-like subunits would complement the corresponding

defective E. coli FI subunits. The 6 and E subunits from

both chloroplast and cyanobacteria proved capable of

partially substituting for the E. coli 6 and e subunits.

The a subunit defective E. coli strain was also complemented

with both the chloroplast and cyanobacteria a subunits, but

to a lesser extent than the 6 and e subunits. The

cyanobacterial P subunit partially substituted for E. coli 3

subunit, but the 0 subunit from chloroplast failed to

complement. The y subunit defective E. coli strain was not

complemented by either chloroplast or cyanobacteria 7

subunits. The ability of F, subunits from chloroplast or

cyanobacteria to substitute for the E. coli F, subunits

demonstrated the remarkable conservation of structure and

function between FIFo ATP synthases.

Biochemical Similarities

F, and Fo sectors can be dissociated with low ionic

buffers freeing the F, sector to become a soluble ATPase,

and leaving the Fo sector in the lipid bilayer as a

functional proton pore. This has facilitated much of the

biochemical analysis of the FIFo ATP synthase.

F, ATPase is composed of a catalytic core of a3#37

subunits that are found in all F-type ATPases. The

catalytic core is sufficient for ATP hydrolysis activity.

ATP synthases contain at least two other subunits, which are

necessary for coupling proton translocation to ATP

production. The e subunit appears to have an inhibitory

action on F, ATPase activity (Klionsky et al., 1984; Dunn et

al., 1987). The inhibition resulting from the e subunit is

apparently reversed by addition of the detergent

lauryldimethylamine oxide (LDAO) (Ldtscher et al., 1984).

The 6 subunit (or the OSCP subunit in mitochondria) is

required for the binding of the F, sector to the Fo sector

(Bragg et al., 1973; Sternweis & Smith, 1977).

The kinetics of nucleotide binding and catalysis have

been extensively investigated (for reviews see Futai, et

al., 1989; Senoir, 1990; Cross, 1992; Boyer, 1993). F, has

multiple nucleotide binding sites three of which are thought

to be catalytic and three other binding sites serving in a

non-catalytic capacity. The catalytic sites undergo uni-

site and multi-site catalysis depending on the nucleotide

concentration. Uni-site catalysis requires that ATP be

present in nM concentrations, or less than the concentration

of enzyme. Multi-site catalysis occurs at AM concentrations

of nucleotide and displays cooperativity (Boyer, 1993).

Since cellular levels of ATP are in the 2-4 mM range in E.

coli (Kashket, 1982) multi-site catalysis probably reflects

enzyme function in vivo. The proposed catalytic nucleotide

binding site in the f subunit contains a highly conserved

amino acid sequence found in other nucleotide binding

proteins such as adenylate kinase and p21 ras (Duncan &

Cross, 1992).

Non-hydrolyzable ATP analogs, high concentrations of

Mg2+, and ADP inhibit F, ATPase activity. F, ATPase activity

is also inhibited by its binding to Fo at pH 7, and this

observation has been attributed to the coupling mechanism of

ATP catalysis and proton translocation. The inhibition

caused by coupling can be reversed by raising the pH to 9.1

(Cain & Simoni, 1989). However, the increased F, ATPase

activity is still sensitive to dicyclohexylcarbodiimide

(DCCD) (see below). This indicates that pH 9.1 and DCCD

affect the ATPase activity of the F, sector bound to the Fo

sector by different mechanisms.

The Fo sector in FFo ATP synthase invariably contains a

small proteolipid c-like subunit that binds DCCD. DCCD

covalently modifies a carboxyl acid in the c-like subunit

(ca, in E. coli) and inhibits proton translocation. DCCD


will also inhibit the activity of Fi ATPase when it is bound

to the Fo sector. The c-like subunit is found in 101

copies in the Fo sector of E. coli (Hermolin & Fillingame,

1989), and believed to exist in multiple copies in other

FFo ATP synthases. ATP synthases also contain a single

copy of an a-like subunit in the Fo sector. Both subunits

have regions of amino acids that are highly conserved

throughout FIFo ATP synthases. Bacterial enzymes contain a

b subunit with the stoichiometry of 2 subunits per Fo sector

which probably form a dimer (Aris & Simoni, 1983; Hermolin

et al., 1983; Dunn, 1992). Comparable b-like subunits have

been found in chloroplast systems, as well as a b-like

subunit in yeast (Velours et al., 1988). No clear b-like

analogs have been identified in mammalian mitochondrial

systems; however, an analogous subunit has been proposed

(Montecucco et al., 1983; Torok & Joshi, 1985; Walker et

al., 1987, 1991; Papa et al., 1992). In E. coli the b

subunit has been demonstrated to be involved in F, binding

to Fo (Hoppe et al., 1983; Perlin et al., 1983) as well as

potentially make up part of the stalk structure (see below).

The Fo sector functions to transport protons across the

membrane. Active proton translocation requires the Fi

sector. Proton conductance occurs passively when the F,

sector has been stripped from the membrane using low ionic

strength conditions. The mechanism of proton transport is

unknown, though models have been proposed involving a proton

wire (Cox et al., 1986, Schneider & Altendorf, 1987), a

proton channel (Deckers-Hebestreit & Altendorf, 1992b) and a

proton binding site (Fillingame et al., 1993). The work of

Dimroth and colleagues in P. modestum (see above) has

elicited the most detailed picture of proton conductance so

far and is outlined below.

An advantage of the P. modestum FIFo ATP synthase is

that the translocation of sodium ions can be measured

directly with "Na+. This was used to look at the action of

P. modestum Fo mediated Na+ conductance (Kluge & Dimroth,

1992). The kinetics of the reconstituted system supports a

transporter mechanism involving at least four steps. The

first proposed step of Na+ conductance was binding of a

hydrated Na+ ion to a defined site on the membrane assumed

to be the Fo sector. In the second step, the Na+/binding-

site complex would 'translocate' to the other surface of the

membrane. Presumably this is through a conformational

change upon the binding of the hydrated Na+ ion. The third

step would be the release of the hydrated Na+ ion from the

binding-site. The last step would be the return of the ion

binding-site back to the original side of the membrane. If

a membrane potential was present then the binding site

returned unloaded. However, if the membrane potential was

absent the ion binding site could return to the original

surface loaded. The flow of Na+ in this experiment was the

equivalent of periplasmic to cytoplasmic, or in the

direction of ATP synthesis. A voltage dependent step was

not found for P. modestum FIFo ATP synthase (Dmitriev et

al., 1993) suggesting a membrane potential is not essential

for the function of the intact complex.

The presence of Na+ in P. modestum FIFo ATP synthase

blocks the translocation of H+/H3O+, apparently by

competitive inhibition (Laubinger & Dimroth, 1989; Laubinger

et al., 1990). Kluge and Dimroth (1992) reported that the

presence of Na+ also blocks the binding of

dicyclohexylcarbodiimide (DCCD), the covalent inhibitor of

Fo mediated ion translocation. DCCD inhibits FIFo ATP

synthase activity by modifying conserved carboxylic acid in

the proteolipid c subunit, (ca, in P. modestum, c,6 in E.

coli). The kinetics of DCCD inactivation revealed a pK(Na+)

of 2.5 to 3.3, depending on the pH. The pH dependence on

Na* binding was consistent with a competition for a ion

binding site. Positive cooperativity (nH=2.6) was also

reported with respect to Na+ transport and Na+-activated

ATPase activity at pH 9.0 (Kluge & Dimroth, 1993). The data

strongly suggest the existence of a ion binding site(s) that

effects the binding of DCCD to the essential amino acid cEa

(c,6 in E. coli). A mechanism involving an essential
arginine was proposed, in which either the amino acid aR,

or cR41 is necessary for deprotonating cB0 so that it can

bind a Na+ ion (Kluge & Dimroth, 1993). This mechanism can


be applied to other FIFo ATP synthases using Ho30 instead of

a hydrated Na+ ion.

Structural Similarities

The FI sector has been extensively studied at the

structural level as well. The structure of the F,-ATPases

from many sources appears to be conserved as assessed by the

resolution obtainable with electron microscopy. Electron

micrographs of several purified F1-ATPases revealed a

roughly spherical shape with a diameter of 90-100 A (Akey et

al., 1983; Tsuprun et al., 1984; Tiedge et al., 1985;

Boekema et al., 1986, 1988; Gogol et al., 1989a, b; Ishii et

al., 1993). The most revealing work so far has been

accomplished by Capaldi and colleagues, and reviewed in

greater detail below. Greater resolution has been obtained

with an X-ray crystal structure of the rat and bovine heart

mitochondria, and is reviewed in greater detail below

(Bianchet et al., 1991; Abrahams et al., 1993).

Cryoelectron microscopy of the FI sector. The most

convincing structural data on E. coli FIFo ATP synthase

comes from Capaldi and colleagues (Gogol et al., 1987,

1989a, 1989b). Cryoelectron microscopy methods have been

used to deduce the structure of the FIFo ATP synthase

complex. The technique involves use of quick freezing of

FIFo ATP synthase to suspend complexes containing liposomes

in a layer of amorphous ice. The data support an

alternating arrangement of a/f subunits arranged in a

hexagonal circle with a diameter ca. 90 A (Gogol et al.,

1989b). An asymmetrical central mass was also observed in

the middle of the a/f ring structure. Immuno-electron

microscopy with anti-a, anti-y, anti-6, and anti-e Fab

antibody fragments was used to investigate the central mass

(Gogol et al., 1989b; Capaldi et al., 1992). The anti-a Fab

fragments were found to bind asymmetrically, allowing the

alignment of each image. The central mass was apparently

composed of the 7, 6 and e subunits, and found to associate

with the f subunits. The positions of the y and e subunits

in relationship to each other were found to be dependent on

ligand binding (Gogol et al., 1990), and correlated with

other ligand dependent studies (see below). The work of

Capaldi and colleagues supports an asymmetrical model with

an alternating hexagon of a/f subunits that produces a

central cavity partially obstructed with the other Fi


Electron crystallography has also been used on F,-

ATPase crystals from the thermophilic bacteria strain PS3

(Ishii et al., 1993). The molecular dimensions and

structure observed in the FI sector from strain PS3 were

similar to those reported for the E. coli F, sector (see

above). The central mass was localized to the lower half of

the central cavity.

X-ray crvstalloQraphy of the FI sector. The X-ray

crystal structure of bovine heart mitochondria F, ATPase has


been reported by Abrahams et al. (1993) at 6.5 A resolution.

F, appeared to be roughly spherical, 110 A in diameter, and

this correlated with electron microscopy data (see above).

However, additional novel new features were also reported.

A 40 A stem composed of two a-helices in a coiled-coil

conformation extended from the globular sphere. One of the

a-helices in the stem extends 90 A into the center of the

enzyme and emerges into a 15 A dimple on the other side of

the globule. The stem is proposed to be part of the stalk

domain that connects F, to Fo, but may also be part of the

central mass observed in cryoelectron microscopy (see above)

since one of the a-helices extends through the center of the

a/f hexagonal ring. A pit that extends 35 A into the F,

particle was observed next to the stem region and proposed

to be occupied by either OSCP or the b subunit. The

asymmetry demonstrated by the crystal of the bovine F,-

ATPase is in marked contrast to the symmetry reported for

the rat Fi-ATPase by Amzel et al. (1992). The cryoelectron

microscopy data supports the asymmetrical structure, calling

the symmetrical rat structure into considerable dispute.

The stalk structure. Side views of the intact

membrane-bound E. coli FIFo ATP synthase complex by

cryoelectron microscopy revealed a stalk structure

connecting the F, sector to the membrane (Licken et al.,

1990). The hexagonal F, complex appeared elongated in

profile images (110 A by 90 A). Equal numbers of two images

appeared in these experiments, a bi-lobed image and a tri-

lobed image. This data also revealed the presence of a

stalk region connecting F, and Fo. The stalk was observed

to be approximately 40 A in length and 25-30 A in diameter.

Fo appeared to have a small protrusion from the cytoplasmic

(F,) side of the membrane, but seemed relatively flat on the

periplasmic facing (60 A in the transmembrane direction).

A 40 A long stem, reported to be in a left-handed

coiled-coil conformation, is also observed in the X-ray

crystal structure (Abrahams et al., 1993). The existence of

a stem, and the corresponding pit, indicates a potential

conformation for the stalk structure. However, the crystal

structure was composed solely of F, subunits, so it is

uncertain whether the stem is an artifact due to the absence

of OSCP or Fo.

The stalk is proposed to be composed of one or more

subunits from F, and in the case of bacterial and

chloroplast systems one or more subunits from Fo. Subunits

forming a proposed stalk region are hypothesized to play a

primary role in coupling proton translocation to ATP

synthesis. Therefore, mutations in stalk subunits might be

expected to alter the coupling mechanism. Mutations

affecting the coupling mechanism have been reported in two

stalk subunits, 7 (Shin et al., 1992) and b (McCormick et

al., 1993).

However, the existence of an extended stalk remains

controversial. The f subunit of F, has been reported to

form a disulfide bond with the b subunit of Fo (Aris &

Simoni, 1983), in which b0l is the only cysteine in the b

subunit and is almost certainly located near the membrane

(Schneider & Altendorf, 1985). Evidence against the stalk

structure also includes the observation that binding of the

F, sector to the Fo sector protects the b subunit from

proteolysis and antibody binding (Hermolin et al., 1983;

Hoppe et al., 1983; Perlin et al., 1983; Perlin & Senior,

1985; Deckers-Hebestreit & Altendorf, 1986; Deckers-

Hebestreit et al., 1992). Additional evidence against an

extended stalk structure was reported by Aggeler et al.

(1987). The protein reactive reagents 4-[3H] formylphenyl

phosphate (HFPP) and 1,2-[3H] dipalmitoyl-sn-glycerol 3-


(arylazidePE) are amphipathic and associate with the lipid

bilayer like phospholipids. HFPP and arylazidePE react with

proteins that come in close proximity to the phospholipid

head groups of the membrane bilayer. The 6 subunit was

protected from labeling by HFPP and arylazidePE only when

the F, sector was bound to the Fo sector. The work on the

b and 6 subunits suggests that they may be buried in the

interface of the F, and Fo sectors, and not exposed as would

be predicted in the extended stalk hypothesis. Mutations in

the c subunit gene from E. coli subunit gene have been

reported to affect the binding of the F, sector to the Fo

sector (Mosher et al., 1985; Miller et al., 1989; Fraga &

Fillingame, 1989, 1991; Deckers-Hebestreit & Altendorf,

1992a; Fillingame 1992a, b, 1993). However, the possibility

that the c subunit mutants effect a potential stalk protein

like the b or 6 subunits can not be excluded.

A model of E. coli FIFo ATP synthase is presented in

Figure 1-1. The model is adapted from Papa et al. (1992).

Proposed positions in the complex for each of the subunits

is based on various sources of data discussed throughout

this Chapter.

Binding Change Mechanism

The binding change mechanism was developed to explain

unique features of ATP catalysis in FIFo ATP synthase.

Unlike all other ATPases that have been studied, F,-ATPases

(and similarly V-type ATPases) do not go through a

phosphorylated-enzyme intermediate. Kinetic studies have

shown that the formation and hydrolysis of the

phosphodiester bond is almost at energetic equilibrium.

Therefore, the energy requiring steps for ATP synthesis is

not bond formation, but binding of Pi and release of ATP

from the catalytic site (reviewed in Senior, 1990).

The binding change mechanism proposed that synthesis of

ATP is accomplished through changing ligand affinities in


Figure 1-1. Diagram of the E. coli FIFo ATP synthase.
Model is adapted from Papa et al. (1992).


the catalytic site of FIFo ATP synthase. Ligand affinities

are altered through changes in protein conformation.

Conformational changes are induced by the proton motive

force, and the translocation of protons. Cooperativity also

plays a role in the necessary conformational changes for ATP

synthesis, in which binding of ADP + Pi at a second

catalytic-site appears to be required. The translocation

of protons by ATP hydrolysis is thought to occur via a

reversal of this mechanism.

Evidence for the binding change mechanism has been

reviewed in great detail by Boyer (1993). The evidence for

the binding change mechanism can be divided into three

categories; nucleotide cooperativity, conformational

changes, and asymmetrical features. Nucleotide

cooperativity has been seen with both hydrolysis and

synthesis of ATP (Matsuno-Yagi & Hatefi, 1985; Stroop &

Boyer, 1985; Cunnigham & Cross, 1988). Conformational

changes due to membrane energization have been seen

indirectly with changes in binding affinities (Chernyak &

Kozlov, 1979; Penefsky, 1985; Matsuno-Yagi et al., 1985;

Lunardi et al., 1988; Zhou & Boyer, 1993), and

conformational changes have been seen directly by exchange

of hydrogens during catalysis (Ryrie & Jagendorf, 1972; Du &

Boyer, 1989). Conformational changes due to ligand binding

have been reported for differential cross-linking (Mendel-

Hartvig & Capaldi, 1991a, b; Aggeler et al., 1992) and

electron microscopy (Gogol et al., 1990). Asymmetrical

features have also been observed indirectly (Grubmeyer &

Penefsky, 1981; Nalin et al., 1985) and directly (Capaldi et

al., 1992, Abrahams et al., 1993). A model of the binding

change mechanism is presented in Figure 1-2 (Boyer, 1993).

E. coli as a Model for Other ATP Svnthases

Attention has recently focused on a type of genetic

disease which is maternally inherited and localized to the

mitochondrial genome (mtDNA) (Wallace, 1992a, b). These

abnormalities produce defects in OXPHOS, and are often

referred to as mitochondrial myopathies. The severity of

this type of genetic disease will depend on the ratio of

defective to functional mitochondria inherited from the

mother. Maternal inheritance of mtDNA is greater than 99%

(Gyllensten et al., 1991) and a chief indicator for

mitochondrial genetic diseases. The heterogeneous nature of

mitochondria distribution to daughter cells results in an

systematic expression of the genetic disease in differing

tissue types (Ashley et al., 1989). Organ and tissue

systems, such as brain, muscle, heart, kidney, liver and

pancreatic islets which utilize OXPHOS as their primary

energy source show a greater sensitivity to these genetic

diseases (Wallace, 1992a, b).

Step 3


Step S

Figure 1-2. Model of the binding change mechanism and
its relationship to energy supplied by proton translocation.
Modified figure from Boyer (1993).

Mitochondrial genetic diseases display many clinical

symptoms that correlate with damage in the organ and tissues

systems listed above. The manifestations of mitochondrial

genetic diseases include blindness, deafness, dementia,

movement disorders, weakness, cardiac failure, diabetes,

renal disfunction, and liver disease (Wallace, 1992a, b).

Clinically the diseases usually present as a degenerative

process with a progressive deterioration from onset. The

degenerative nature of these genetic diseases can be

attributed to a gradual loss of mitochondrial function that

occurs over a person's lifetime. The process of OXPHOS is

believed to contribute to the gradual loss of mitochondrial

function through damaging mtDNA with oxygen free radicals.

The OXPHOS pathway is composed of five enzyme complexes

containing multiple subunits. The first four enzyme

complexes constitute the electron transport chain. NADH

dehydrogenase (complex I), ubiquinol-cytochrome c

oxidoreductase (complex III), and cytochrome c oxidase

(complex IV) have subunits encoded in the mtDNA as well as

the nuclear genome. Succinate dehydrogenase (complex II) is

encoded solely in the nuclear genome and is not implicated

in maternally inherited mitochondrial genetic diseases. The

electron transport chain produces a proton electrochemical

gradient which is utilized by FIFo ATP synthase (complex V)

to produce ATP.

Various abnormalities in mtDNA have been observed

associated with mitochondrial genetic diseases. Deletions

and insertions in the mtDNA genome have been associated with

Kearns-Sayre syndrome (KSS), chronic external

ophthalmoplegia plus (CEOP), and Pearson's marrow/pancreas

syndrome. The mitochondrial genetic diseases myoclonic

epilepsy and ragged-red fiber disease (MERRF); mitochondrial

encephalomyopathy, lactic acidosis, and stroke-like symptoms

(MELAS); and maternally inherited myopathy and

cardiomyopathy (MMC) have been associated with point

mutations in mitochondrial encoded tRNAs. The MERRF, MELAS

and MMC mutations are postulated to act by inhibiting

mitochondrial protein synthesis. Deletions, insertions, and

point mutations in mitochondrial tRNAs have the effect of

reducing or eliminating all or several of the proteins

necessary for OXPHOS. This global effect on OXPHOS often

results in morphologically abnormal mitochondria, and some

of the more clinically severe manifestations of the

mitochondrial genetic diseases (Wallace, 1992a, b).

Point mutations in mtDNA also result in mitochondrial

genetic diseases by producing missense mutations in

mitochondrial genes which impair their function. One

example of this type of mitochondrial disease occurs in

Leber's Hereditary Optic Neuropathy (LHON). Initially LHON

was linked to a missense mutation in the gene for the ND4

protein which converted amino acid 340 from a conserved

arginine to a histidine. Additionally eight separate

missense mutations have also been linked to LHON. The other

missense mutations linked to LHON occur in other

mitochondrial encoded subunits of NADH dehydrogenase

(complex I) as well as one missense mutation in the cyt b

gene, the mitochondrial encoded subunit for

ubiquinol:cytochrome c oxidoreductase (complex III) (Singh

et al., 1989; Huoponen et al., 1991; Howell et al., 1991;

Brown et al., 1992).

The second genetic disease to be linked to a

mitochondrial missense mutation is neurogenic muscle

weakness, ataxia, and retinitis pigmentosa (NARP) (Holt et

al., 1990). The missense mutation is at nucleotide 8993 in

the mtDNA ATPase 6 gene. The mutation results in a change

in ATPase 6 amino acid 156 converting a conserved leucine to

an arginine (ATPase 6uj). The mutation also resulted in a

restriction fragment length polymorphism (RFLP) resulting in

the appearance of an additional Aval site in the mtDNA

genome. The Aval RFLP allowed the correlation between the

severity of the disease and the percentage of mutant

mtDNA genomes present in each patient. The mutation could

also be followed by a HpaII RFLP.

The NARP mutation has also been linked to a second

mitochondrial genetic disease. The mutation has been linked

to some cases of maternally inherited Leigh's syndrome, or

subacute necrotizing encephalopathy (SNE), by Aval and HpaII

RFLP analysis (Tatuch et al., 1992; Shoffner et al., 1992;

Ciafaloni et al., 1993; Yoshinaga et al., 1993). Leigh's

syndrome is an early onset encephalopathy that often results

in death at a young age. Leigh's syndrome was apparent only

in patients with very high levels of the NARP mutation. The

relationship between these genetic diseases is shown by the

fact that the patient from the NARP pedigree that had very

high levels of the mutation (>95%) exhibited clinical

symptoms of Leigh's syndrome (Holt et al., 1990).

Additionally a patient in one of the Leigh's syndrome

pedigrees with lower levels of the mutation present (<80%),

exhibited clinical symptoms of NARP (Tatuch et al., 1992).

A second missense mutation at L156 has been reported by

de Vries et al. (1993). The nucleotide mutation T8993-C was

found instead of the T-G changed previously observed. The

resulting mutation ATPase 6,L5p could be followed by HpaII

RFLP, but the Aval RFLP was not present. The T8993-G

mutation results in both restriction sites being formed,

while T8993-C results in a HpaII site, shifted one base pair

downstream from the T-G HpaII site. The NARP mutation

reported by Yoshinaga et al. (1993) was tested only by HpaII

RFLP, so a question exists as to which mutation is present.

Clinically the patients with NARP did not always

exhibit elevated levels of lactic acid which supports the

premise of a functional electron transport chain in some of

the patients (Holt et al., 1990). Neurological atrophy was

the prevalent phenotype of the disease indicating that the

cellular energy thresholds were not being met in this tissue

type. A defective ATP synthase (complex V) could account

for this phenotype. Attempts to analyze cultured cell lines

from patients gave results indicating no effect (Tatuch, et

al., 1992), or only a partial loss of FIFo ATP synthase

function (Tatuch & Robinson, 1993). Determination of the

probable biochemical affect of the NARP mutations is

presented in Chapter 5.

E. coli FiF_ ATP Svnthase


E. coli FIFo ATP synthase is well characterized

genetically. FIFo is composed of eight different subunits

in the following stoichiometry: a(3), 8(3), y(l), 6(1), and

e(1) in F,, and a(l), b(2), and c(10l) in Fo. All subunits

are encoded in the unc operon in E. coli. The order of

genes is: uncB (a), uncE (c), uncF (b), uncH (6), uncA (a),

uncG (7), uncD (0), and uncC (e). An open reading frame

termed uncI also resides 5' with respect to uncB (a). The

uncI open reading frame does not encode a known subunit of

FFo ATP synthase. The uncI gene encodes a polypeptide

which has been translated in vitro (Schneppe et al., 1991).

However, mRNA studies indicate that the RNA encoding uncI is

quickly degraded from the polycistronic unc operon message,

and the rate of mRNA degradation may play a role in the

differing protein levels of the FIFo ATP synthase subunits

(Schaefer et al., 1989; Lagoni et al., 1993). Differential

translational control has been demonstrated for the unc

operon message based on ribosome 'toe prints' (Schaefer et

al., 1989).

F1 Subunits

The FI sector of ATP synthase houses the catalytic

activity of the enzyme as discussed above. All five of the

FI subunits, a, 0, 7, 6, and e, are required for both ATP

synthesis and ATP-driven proton translocation in vitro (Dunn

& Heppel, 1981) and in vivo (Downie et al., 1979; Humbert et

al., 1983).

The a subunit. The a subunit is 513 amino acids long

and has a deduced molecular weight of 55,300 daltons. It

folds into a roughly ellipsoidal, globular structure (Gogol

et al., 1989a; Ishii et al., 1993;Bianchet et al., 1991).

The a subunit can be divided into three functional regions,

the membrane-binding region, the nucleotide-binding region,

and the a/# signal transmission region (Senior, 1990). The

membrane-binding region resides in the amino terminus of the

a subunit, and may be involved in the association of the 6

subunit with Fi. The depletion of the 6 subunit from the F,

sector results in the loss of the ability of Fi to bind Fo.

The function of this region is supported by proteolysis

studies (Dunn et al., 1980) and mutagenesis studies (Maggio

et al., 1988). The nucleotide-binding-domain is between

amino acid positions aE and ao and proposed to be involved

in the tight binding of ATP at a non-catalytic site. The

nucleotide-binding domain function is supported by sequence

homology with other nucleotide-binding proteins (Maggio et

al., 1987) and mutagenesis (Rao et al., 1988). The proposed

a a/# signal transmission region is believed to be involved

in transmission of cooperativity between the catalytic

nucleotide-binding sites. The a/f signal transmission

region is located at amino acids positions a34 to a375 and its

function supported by mutagenesis experiments (reviewed in

Senior, 1990).

The B subunit. The # subunit is slightly smaller than

the a subunit with 459 amino acids and a deduced molecular

weight of 50,200 daltons. The # subunit appeared to be very

similar to the a subunit in structural studies (Gogol et

al., 1989a; Ishii et al., 1993; Bianchet et al., 1991), and

shares primary amino acid sequence homology with the a

subunit (Walker et al., 1984). The catalytic nucleotide-

binding domain believed to be contained in the f subunit

between amino acid positions (3o and f833 (Duncan et al.,

1986; Senior, 1988). Support for the # subunit containing

the catalytic site of F, comes from affinity labeling with

nucleotide analogs (reviewed by Senior, 1988) and

mutagenesis analysis (reviewed by Futai et al., 1989;

Senior, 1990; Schnizer, 1993).

The 7 subunit. The 7 subunit consists of 286 amino

acids with a deduced molecular weight of 31,400 daltons.

The 7 subunit displays conservation at the amino acid level

in the amino-terminal and carboxyl-terminal regions (Futai

et al., 1989; Iwamoto et al., 1990; Shin et al., 1992). The

7 subunit has also been localized to the central cavity of

the a/S ring in F, by immuno-cryoelectron microscopy (Gogol

et al., 1989). The amino-terminal portion of the 7 subunit

was shown to reside in the cavity by using cryoelectron

microscopy of monomaleimidonagold cross-linked to cysteines

that had been incorporated into the 7 subunit via site-

directed mutagenesis of the uncG (7) gene (Wilkens &

Capaldi, 1992). Abrahams et al. (1993) proposed that the 7

subunit may be responsible for a 90 A long alpha-helix

extending through the central cavity of F, (see above).

Aggeler and Capaldi (1992) also reported cross-linking

via tetrafluorophenylazide maleimides (TFPAMs) of the a and

f subunits to the mutant 7 subunits containing cysteine

substitutions. The mutant 7syc subunit could be cross-

linked to the f subunit, yielding different products with

different ligands present (see above). The mutant 7v6c

subunit cross-linked with the a subunit in an non-substrate

dependent manner. Both cross-links inhibited FI-ATPase

activity at levels proportional to the cross-linking yield.

The cross-linking results correlate with mutagenesis

analysis of the 7 subunit. The carboxyl-terminal region of


the y was found to be important for ATPase activity (Iwamoto

et al., 1990), in agreement with the 7'yTc/a cross-link.

Substitutions of arginine and lysine at 7'y affect coupling

between catalysis and proton translocation (Shin et al.,

1992), in agreement with ligand dependent 7s c/f cross-

links. Apparent ligand induced conformational changes

indicates that the y subunit may play a role similar to the

e subunit (see below).

The 6 subunit. The 6 subunit is an elongated and

highly a-helical subunit of 177 amino acids with a deduced

molecular weight of 19,300 daltons (Sternweis & Smith,

1977). The 6 subunit is required for F, binding to Fo

(Bragg et al., 1973; Sternweis & Smith, 1977). The 6

subunit can be cross-linked to the a and f subunits via the

reactive cysteine Sc,4 (Aggeler & Capaldi, 1992). The 6-a

cross-link does not result in loss of Fi-ATPase activity

(Brag & Hou, 1986; Tozer & Dunn, 1986; Mendel-Hartvig &

Capaldi, 1991a). The amino-terminal and carboxyl-terminal

regions of the 8 subunit have been reported to be oriented

toward F, (Mendel-Hartvig & Capaldi, 1991a).

The 6 subunit has been shown to be involved in F,

binding to Fo by mutagenesis analysis (Humbert et al.,

1983). Mutations in the 6 subunit have also been implicated

in affecting the stability of the FIFo ATP synthase complex

(Stack & Cain, 1994) and DCCD sensitivity (Hazard & Senior,

1994). The 6 subunit has also been suggested to open the

proton channel by interacting with Fo (Angov et al., 1991).

The e subunit. The e subunit is a globular protein

(Sternweis & Smith, 1980) of 138 amino acids with a deduced

molecular weight of 15,000 daltons. Although structural

data has not yet been reported, the E subunit has been

crystallized separately and in conjunction with the 7

subunit (Codd et al., 1992; Cox et al., 1993). Binding of

the e subunit to FI-ATPase (with or without the 6 subunit

present) results in inhibition of ATP hydrolysis activity

(Laget & Smith, 1979; Sternweis & Smith, 1980; Dunn et al.,

1987). The e subunit is also required for F, binding to Fo

(Sternweis, 1978). The two functional domains of the e

subunit have localized to E1lg for F, binding to Fo, and 09

for inhibition of F,-ATPase activity (Kuki et al., 1988).

As mentioned above the detergent LDAO has also been proposed

to effect the inhibitory action of the e subunit (L8tscher

et al., 1984).

The e subunit has also been reported by Capaldi and

colleagues to undergo conformational changes that affect the

subunit's rate of cleavage by trypsin, which results in

activation of FI-ATPase activity. The conformational change

was controlled by the binding of ligands to F,. Ligands ATP

+ EDTA, ADP + EDTA, AMP-PNP + Mg2+ resulted in a fast rate

of trypsin cleavage. Ligands ATP + Mg2+ or ADP + Pi + Mg2+

resulted in a slow rate of trypsin cleavage. Nucleotide

binding was required to elicit the conformation with a fast

rate of cleavage (EDTA, Mg2+, Mg2+ + Pi, EDTA + Pi were

equivalent to ATP + Mg2+). The rate of trypsin cleavage

with the ligands ADP and Mg2+ was high, but slowed by

addition of phosphate (Pi). The concentration of Pi was

determined to change the conformation of e in a manner that

correlated with high affinity binding of Pi to F, (Mendel-

Hartvig & Capaldi, 1991a). The indication from the

different ligands analysized is that the conformational

change depends on whether ADP or ADP + Pi is present in the

active site. This assumes Mg2+ is required for Pi binding,

and that AMP-PNP mimics ADP in the catalytic site. The e

subunit could also be locked in either conformation by

treatment of FIFo ATP synthase with DCCD under the proper

ligand conditions (Mendel-Hartvig & Capaldi, 1991b). The

ligand induced conformational change of the e subunit could

also be followed by immuno-cryoelectron microscopy (Gogol et

al., 1990), as well as, differing cross-link product yields

(Mendel-Hartvig & Capaldi, 1991a; Aggeler et al., 1992).

The e subunit has been cross-linked via 1-ethyl-3-[3-

(dimethylamino)propyl]-carbodiimide (EDC) to f (Mendel-

Hartvig & Capaldi, 1991a), and cysteine substitution mutants

to a and 7 (Aggeler et al., 1992).

The FE Subunits

The Fo sector is made up of three subunits in E. coll.

All three subunits, a, b and c, are required for coupled

proton translocation (Schneider & Altendorf, 1985). The


stoichiometry of subunits is ab2c10,o (Hermolin & Fillingame,

1989). As mentioned above, E. coli Fo is insensitive to

oligomycin. In E. coli, proton translocation is usually

inhibited by covalent modification of cD1 with DCCD.

The a subunit. The a subunit is a hydrophobic protein

of 271 amino acids and deduced molecular weight of 30,276

daltons. However, the hydrophobic nature of the a subunit

is thought to cause anomalous binding of sodium dodecyl

sulfate (SDS), resulting in an apparent molecular weight

around 23,000 daltons by SDS polyacrylamide gel

electrophoresis (SDS PAGE). Hydropathy plots of the a

subunit predict four-five potential membrane spanning

regions. Predictions for the number of membrane spanning

regions range from 5-8, and are discussed in greater detail

below. Truncations of the a subunit result in loss of FFo

ATP synthase function if more than seven amino acids are

removed from the carboxyl terminus (Eya et al., 1991).

However, missense mutations can also abolish enzyme function

(Cain & Simoni, 1986).

The a subunit contains two regions strongly conserved

at the amino acid level (see Figure 1-3). These regions of

homology have been the basis for extensive site-directed

mutagenesis analysis (Cain & Simoni, 1988, 1989; Eya et al.,

1988, 1991; Howitt et al., 1988, 1993; Lightowlers et al.,

1987, 1988; Vik et al., 1988, 1990, 1991, 1994; Chapter 4

and 6). An overview of the results of these studies is

presented in Chapter 7. The studies indicate that the a

subunit is involved in proton translocation and coupling of

the proton motive force to ATP catalysis. The amino acid

ao is considered to be essential in proton conductance,

with amino acids aE,1 and agm, also playing important roles.

The b subunit. The b subunit is a largely hydrophilic

protein of 156 amino acids and a deduced molecular weight of

17,265 daltons. The b subunit is anchored to the membrane

by one predicted membrane spanning region in its amino-

terminal end. When this region is removed via a deletion in

the uncF (b) gene, the hydrophilic portion of the b subunit,

b,, is found to form a dimer (Dunn, 1992). The hydrophilic

portion of the b subunit has been shown to play an important

role in F, binding by proteolysis experiments (Hoppe et al.,

1983; Perlin et al., 1983).

Mutagenesis experiments have resulted in mainly

assembly phenotypes (Jans et al., 1984, 1985; Porter et al.,

1985; McCormick & Cain, 1991). Mutations at bA, result in

assembly mutations, with an apparent dependence on

expression levels (McCormick et al., 1993). Additionally

the bAK mutation results in a unique phenotype that effects

coupling of proton translocation to ATP catalysis. This

mutation had a severe effect on ATP synthesis while

maintaining substantial ability to actively transport

protons with ATP hydrolysis. This is the clearest reported

example of an asymmetrical effect on the coupling mechanism,

indicating that the b subunit plays a direct role in the

mechanism. For this reason it is very interesting that

mammalian mitochondria do not contain an analogous subunit.

This indicates either a major change in the mechanism of

coupling or a major change in the b subunit.

The c subunit. The proteolipid c subunit is an

extremely hydrophobic protein of 79 amino acids with a

deduced molecular weight of 8,288 daltons. It folds into a

hairpin of two a-helices that span the membrane. A polar

loop connects the two a-helices, and has been demonstrated

to extend from the membrane on the cytoplasmic side (Girvin

et al., 1989; Hensel et al., 1990). The structure of the c

subunit has also been partially solved by NMR in a

chloroform-methanol-water mixture (Girvin & Fillingame,

1993). (For c subunit reviews see Fillingame 1992a, b.)

As mentioned above DCCD binds to c61 and inhibits

proton conductance and the ATPase activity of F, bound to

Fo. The amino acid c,6 is proposed to be near the center of

the lipid bilayer, at least 10 A inside the membrane surface

(Mitra & Hammes, 1990). When site-directed mutagenesis is

used to change c,6 to either a glycine or an asparagine,

proton conductance is abolished. The carboxylic group of

amino acid c,6 is considered essential for proton

conductance, since activity is retained with a glutamate

substitution (Miller et al., 1990). The carboxylic group

can also be moved to the other a-helix of the c subunit,

with retention of some activity in the double mutation

cA4-.D, (Miller et al., 1990). Third-site suppressor

mutants have also been isolated that improve the activity of

the cA24,, double mutation (see below).

The several amino acids located in the polar loop of

the c subunit and are conserved in other proteolipid

subunits. Structurally non-conservative amino acid

substitutions at these positions result in a loss of ATP

synthase function. The phenotype of these mutants is

proposed to be the uncoupling of ATP catalysis from proton

translocation (Mosher et al., 1985; Miller et al., 1989;

Fraga & Fillingame, 1989, 1991). A reduction in F, binding

affinity to Fo was also observed with some of the mutations,

as well as, passive proton conductance when F, was

apparently bound to Fo. Loss of F,-ATPase DCCD sensitivity

was also observed, while proton conductance was still

inhibited by DCCD. The highly conserved amino acid

positions cR41, ce, and cp, were the most sensitive to

mutagenesis (Fraga & Fillingame, 1991).

The c subunit also has two apparent roles, the

conductance of protons and the coupling of F, to Fo. The

mechanism of c subunit coupling appears to be mainly through

the binding affinity of F, for Fo. The a subunit's effect

on binding affinity would have to be mediated through

another subunit if the hypothetical stalk region exist.

Candidates for interaction with the c subunit are the b and

6 subunits.

Second-Site Suppressor Mutations in F0

A second-site suppressor mutation occurs when the

original deleterious mutation is complemented by a second

mutation in another codon. Second-site suppressor mutations

can reveal interactions between amino acids, potentially

giving insight into tertiary structure if the second-site

suppressor resides at a distance from the original mutation.

If the suppressor occurs in another protein, then subunit

interactions may also be revealed. Several suppressor

mutation studies have been carried out in Fo subunits and

are reviewed below.

Kumamoto and Simoni (1986a, 1986b) studied suppression

of the mutation of bGsD. The substitution did not

completely block proton translocation but function of the

complex was so low that the bacteria failed to grow on

succinate minimal medium. A selection scheme was used to

isolate suppressor mutations that occurred in other FFo ATP

synthase subunits. Suppressor mutations occurred in each of

the other two Fo subunits. Two suppressor mutations

occurred in the a subunit, apA and ap2.L The c subunit

suppressor mutation was cA62. Both suppressor mutations

recovered only partial activity for FIFo ATP synthase.

Biochemical analysis of the bG.D mutant, and the nature of

the suppressor mutants, indicated that the suppressor

mutants probably were correcting conformational changes

imposed by the initial mutant. The locations of the

suppressor mutants imply that the b subunit interacts with

the a subunit and the c subunit.

Cain and Simoni (1988) engineered an intra-gene

suppressor mutation in the a subunit gene. This study

involved the two amino acids a,19 and a... Both amino acids

were changed independently and together by site-directed

mutagenesis. The mutant aEH had little activity while the

mutant as, had somewhat higher levels of activity. The

second-site suppression was constructed by the production of

the double mutation of aEo.HE. The strain carrying the

double mutation exhibited more FIFo ATP synthase activity

than either of the individual mutants. The amino acid

positions a1, and a,, are predicted to be on different

membrane spanning segments (see below). The ability to

exchange the amino acids at these two positions indicates

not only interaction between the amino acids themselves, but

interactions between the membrane spanning regions.

Miller et al. (1990) also isolated an internal

second-site suppressor by investigating suppressors of the

c6,., mutation. The c6., mutation abolishes proton

conductance and renders FIFo ATPase insensitive to DCCD

inhibition (see above). Attempts to move the aspartic acid

to amino acid positions near c6, failed to restore enzyme

function, indicating precise requirements for the position

of carboxylic acid in enzyme function. A second-site

suppressor mutation was isolated through limited medium

selection. The cA mutation was found to suppress the

cW1, phenotype, allowing 60% of normal enzyme function.

Interaction between c4 and cq1 had been demonstrated before

with the DCCD insensitive mutant cA.s (Fillingame et al.,

1991). The cA24o. double mutation has been used to align

the positions of the two membrane spanning helices predicted

in the c subunit. The ability to move the essential

carboxylic acid to the other helix of the c subunit

indicates that position cu is in close proximity to position

Cg6. This conclusion has been supported by NMR structural

data for the c subunit (Girvin & Fillingame, 1993).

The cA24D. double mutation was also subjected to a

limited medium selection process (Fraga et al., 1994).

Third-site suppressor mutants were found in the c and a

subunits. Five mutations were found in the c subunit at

amino acid positions cFS3 (3), CM57 (1) and cG, (1). Thirteen

mutations were found in the a subunit at amino acid

positions aD2 (1), a, (2), aA21 (3), a22, (3), and au (4).

There were no clear trends in the type of amino acid

substitutions isolated. Surprisingly, several of the

suppressor mutations occurred in regions predicted to be

outside the membrane hydrophobic core (see below). One

possible explanation for this observation is that extra-

membrane loops may play a role in the positioning of the

membrane spanning regions. This is an interesting

hypothesis, since the majority of the third-site mutations

in the a subunit are in positions that might affect the

positioning of the membrane spanning region containing the

essential amino acid a,0.

The a Subunit of E. coli FIF4 ATP Synthase

Conservation of the a Subunit

The a subunit shares homology at the amino acid level

with all a-like FIFo ATP synthase subunits. This homology is

most prevalent in the carboxyl-terminal region of the subunit.

This region has also been found to play an important role in

proton conductance (Cain & Simoni, 1988, 1989; Eya et al.,

1988, 1991; Howitt et al., 1988; Lightowlers et al., 1987,

1988; Vik et al., 1990, 1991).

Figure 1-3 aligns the reported a-like amino acid

sequences. Six-membrane spanning regions were defined based

on the criteria of exclusion of non-conserved charge from a

hypothetical hydrophobic core. Surprisingly, at least one

amino acid was found to be completely conserved in five of the

six predicted membrane spanning regions. Likewise non-

conserved positively charged amino acids (mostly lysines) were

used to define eight of the twelve hydrophobic-hydrophilic

interfaces. The variation in size of one extra-membrane loop

helps define two other interfaces. The other two interfaces

ON 0% 0N M % 0% 0% 0% al 0%
CA oo H4cv-4 r4 0 H H-V44 r4 HHH
S(O o H % w-
* 0% 0 0 0 4 #00 CO O
U0 -.0 Oi W O -W IA P4 Wr.FA 0 to00 0 0 (7O1
oa) o H 4 $ 4 0 0 ( r-4 0 H a0 ) -4
r, U 9#- 4 0 0 -94 ESPE O
* *r 4 9 V *
44U r *"4v p 0) N 4J 4 V VP'-4
d0) 0 4) ) a) a) 0 40 4
V X4 V u '04)FOt aau''u
P-4 0 0% 0
*q 5 *C *C ** >-
4V UO I > a) 41 a) 4 0V 4) CV V 4-1 UVi0 43 V1p a)
0 ) ONONOCOU 0 N 0 N-4000
kOOH wu04N 04 Ma404MM

(I (
rl -4

(7\1 ai 0%



Sa C ar
* S *

I 0 .)
0 0 U 4
14 1 f 0
PtP40k t

0O I ( N -
Hr-l r-4 r- CO
* H
43434) .4.14

00 0 0 '0'0

C C Q) 44

4j 0 'O '0 4 k 0
to r4 C C PO %4 -4
S1 0k 00 C 0 1
-4 0 4 0 '0
04 C C x0 C C
0 000 0U 00
$4 .0 -4430 *4 .0
0 UC *0P4 *P4 0
0 U 1 -4 0 4 4 U
r-4 ova 0 -q V 0
S-VC 0e C +V
01: r. ,,x: o
MARC ^-> 5 x
U f 0 t U
m 000 q 4U 3 0 C
C ooO P :3 04 0 V 0
4 a Q (B I *H R
O4 4 a -H O Cr 1 gO c

4 0 4 4-4 00 0 0
4) 4 1 C9: Q 0 1` 04 P4 4-B 3 '0 $4 >0 C
c W C A 00 0 U) 0 4) -0 0 .J *C V 0
a o*-4- *- 0 0 t O M'- 0 0E
4 I 3 k I I Q C 41 U 0 0 0 0
0 RF44~(3 M C O 0 0 C C 0 000 0-14 C
3 r. Q Rf-4 0 0 0c -cB C O a0
H 04 5040 N)4 -H4-4-404U-H a *1 r4. 0
N 0 40 0 0U 00 0 Vt14 9 04 0
4 0 & M 0 U 1 Z004 3 0k A04 0 O4O--'0 0 0
4.) 94 *43x .0a) "q t % 0 4.400 QCO 4 0
+> 0U CW 4 0I0 3a O 0AO O 940L5 C U0U '3
H.-1 n MO -A0O,0CUg3 V % 000 C U-*0 0 0
-4 *9 &F 0400 oU009- 4a0>*4 0 U 44 0

o O 03V03 o 0 (4COQ 0 P4 4 4- V0e -H e E
I 1 M 0'* AW4-4H04I gN-*H O- 4CUC440-14
r0 000' >9 00 g-H O4 0-4 S0 H- 4 U 04M 0 W.- 4 >
o BqC P4 A0 >0 U T-O 'r49444 "4 --4 :1 > 0
>C1V04 t U U I z A k

0 U>M N ar04.0 .c-.. NUM0Q4XE U)C3Z0 a>

X Q)
. 0



CM -
*.4 .,-q

to 04L

M 0

S4 M.rO
R0 a

o .
M 040c

a a)
E 0

o) 0
4 4

0 4() '0

0 0 0

N Er- 0

E 0Q

)I 0 c
C 40

*^ Ct>
*i 0 U

************************** 42
*** ********** *

Hu hi==00=' O

a cn toO E-4 N420 zw 000 0maEEU 0

QQWc ow0O w0x 00 80 x

W w ZZ w Zg w W i-0=001 00
H 4> H o N 444a 0>>0 4 > 4 0> 0*Ig42

r2H. 032Y H 4
HH H HH HHHE-4H>>> 5 4 4 0rO4
M1 5 %EMMMx4x4aOHHMMW u 000 I 00 4 -o

2 MH CIooC O zzuI)i RbU-4 Ha zz IE4zz 0
00 M H>aH HHQ

*** ** *** ** ** *** ** 00
*M*QQU **MM *M QQQQQQ tteSBsat< -M

* * *X* *> 0

>M r"-- H >M >< H E> M 2 = HH 44 0Z g04 00
HHHH 0ia 44a0 WH44>>N 04 4 &HHO>H>> 40

0C l ) U O P 4
>>H3H >HHHHH >H ErtCtH >e4 4 E

E C w>wl>< CU 9
con H UZ toZ M a couM 4HQ 0 4

H0 0 H E--4 H E >4 M N N 0 AdW
z W rh Pk l Bx C*0z
lcn IW o n I > O> 9-
fa0 CE C QM !BCE E ME E M &

4 >>u HHH H 0 000 N0 A 04a4 04a ad ad 14(C
cn g l > 0 0 4 Ek o a > 0 i 1 i> 4 *- C A *H H 0 H a
04 *4W 0 in3 0 0 > H 5 CUi >4S N to3NaO CY 01 (n cn(n u
>o ssjH6<> tHHHHO>H Z iZZc XI T4)EPii lQ 9)E-4C X -

PQ>MU044WpHHu HHH>Hicu C O3 U>E43& &.(t.!&rt >> *h ^r VJUE^E^e ^6^|^ > >>0^*-3<|- ^ (O'-<

--- ---------------------------------------------- .

1- -
Q) r
3 R
O 0

0 ,
no w>==> =w>>>>> > =wa rof ga00

z z z 04) 0 4

4 z >zr4 >4 >E4E>4 >4 >4 >4 iC
>> lH

[ > eiCeE-E.Ie1e 00
a000000 Crq****************** 1

0*****C**** *****

* ** ** ** ** ** ** ** ** e*.*
N*c x W* x (Y** **(*Ygo0** C00

>4 SP0 4N I k ** **** *t********va at

0000 000000000000000 0 U


3 A* 04* 04 C C4* 0 000 4 N v

44 N 4^ t ******************** 0 N
XX Si & hjh********.**<** (0(

^M -U<

W^r~~~~ ^^^^^^^^MHIa~hfhhhhM^^M~l~ *j3



H H ,40 4 > > > > >4 H HH4HH N N H 0

00 00000000 0 0

13E 440

M !<



[ >

z -
U (A

41 >.



u u
H >-4

0 **

*c.1H c~U
* J)U)04

:4H HHHHW A 44

.1,4 0 404040400
Z4 0>HHH0004 0
E>> > 0404'4E4UU 1

E-IHft4 >HHHH1-O00A

a4 33013>4 w
<14 0(00 4N
100 I XXXXO >
OW -4 E-4 M


44 N 4 HH
x W En ac W* i

l~~-l HH
0 U
!> >:g
> E] l-2:
04 & AA&





ata4 4& t4 4A

A$i (i 04 4 a {i Ni Ci 0 4 N0 A 04 0' 04 04 N

c 01 ClY a iH ic>1 C4f
U44444C OM
P E a-o E2-4&-4 P i: P

a c H HHH H EfcIN SE-04
Q 010t010101> 4

lHHHHHS> % agg-C


0 (d ia O A 0 9 i0i 0 M U -r A N U i-4 0 k4 0 M m a > 0
W>M>EQ04o0 4 PWoE-0MUZao M 4 X^ E-COU (0 z m =


N ---
N +
N1 H









0-t 0
i) *1


$4 0.

V* O

0 0)
0 0O
x c.4

0 0
0 0

0 e


0 0
e- 0
*- o


aQ rMww 0 WWD 4 w in

4 0 0* z I4 X




00000 00
ooooia aa a




ErJ H BE-1tE"EE-4 E-1 Ert -4>HHE
HHH>HH HH >>> H >>>>>>>>>>>
>>>4 QA 4> 00> >>>><>>>>>>>

29 44U44 0 0 0 0 00
rt!^4><< > HWHE 4. p.HOHHH>>>>>>>>> >>

[* u^ <>^S fcr MIlH "^ M^'-^-^-^(^'-'-
53c'S 5 dsSSHtt^eE'83?0^^^^^^1^^^1^^^^^1


* *



* H

A 0 Q

> K HA
d L (ll -^EL

> 20 f
S> 6<
> u a

* H
*- > H

* 0*
u u
* <
p H H

* **

>4 >



.4, >4 >4 >4

(Q < K

> >
H H > ca *
>>* >
HH >

> Hi
4H E41 >H

13 HrCOH <

SHZ 4 .
L X4 1 l

24 *
0 H )



44 H H H H
4HH00 4-


4 E 0< U
EA OM d M ^


** H


HN 0
H> H U

g < 84 Q

** H |Mi
(S l
*** 1
** V1
*** u

<* **

ri0 to Ol m49 >1 A O M A '-I 'A N 0 4 Oa 04 E0 9 W > 0

r-1 >
m 0



.o u

) i0)
X >




D 0

u0 V

o4) (
0 0

0 0
O T3



0 0

r4 0 0)

0 0

I g o




were defined by more subjective methods. Five of the six

hydrophobic membrane spanning regions were 18 to 24 amino

acids long. Predicted membrane spanning region IV is

defined as only 15 amino acids. This membrane spanning

region does not follow the criteria used to define the other

regions, however its existence is supported by alkaline

phosphatase data (see below).

Additionally, the first extra-membrane loop was also

found to contain amino acid homology. The amino acids am,

aQ6, aE, and aF were found to be conserved in all bacterial

and chloroplast a-like subunits. Predictions of secondary

structure indicate that these amino acids may form an

amphipathic a-helix. Deletion of this extra-membrane region

results in loss of enzyme function (Lewis & Simoni, 1992).

The importance of this loop is also supported by the third-

site suppressor mutations mention above (Fraga & Fillingame,

1994), for one of the mutations isolated was in aD,. The

amino acid a, also appears to be conserved in the

mitochondria of fungi and plants. One possible function of

this region is interaction with the b subunit. This

conservation is lost in the a-like subunits from organisms

that do not have a b-like subunit embedded in the membrane


ToDology of the a Subunit

Models for the topology of the a subunit have predicted

5 to 8 membrane spanning regions. Hydropathy plots of the

E. coli a subunit predict 5 hydrophobic regions large enough

to span the membrane. Vik and Dao (1992) used computer

hydropathy plots from the yeast a-like subunit and the E.

coli a subunit to predict six membrane spanning regions.

The alignment presented in Figure 1-3 is the first to use

all the a-like subunits to determine a consensus for

hydrophobic regions (see above). The alignment data was

used to construct a hypothetical topology model of the a

subunit (Figure 1-4).

Two groups have produced phoA/uncB (alkaline

phosphatase/subunit a) fusions to study the topology of the

a subunit. Bjorbaek et al. (1990) predicts eight membrane

spanning regions with both termini extending into the

cytoplasm. Lewis et al. (1990) also predicts eight spanning

regions but with both termini on the periplasmic side of the

membrane. The production of a phoA fusion at the carboxyl

terminus lends more credibility to the latter model.

Additionally, the Bjorbaek et al. (1990) phoA fusion data

can be more easily incorporated into the Lewis et al. model

than the Lewis et al. (1990) data into the Bjorbaek et al.

model. The Lewis et al. model also predicted transmembrane

spanning regions too short to span a lipid bilayer by the

traditional a-helix. Non-a-helix structure was proposed for

the shorter transmembrane regions. Alternately a structural

shortening of the distance protons must be transported

across the membrane was also proposed. Analysis of the data

L a144 0 0
cn r >ka o: e *e > B
Pm. At 0
0 > e 0 4 0 0-
0 4 1 '0 41 E- M

4 4 E0 -4 4E ) 30

40 C 1. 40

14 'C
4 4

>. E-4 -U
Eet .e "4>
0, 0.


r4 4
u0 a.
0 .5
- E4


1- P
FA r-4
.4-) -p t1 ..l
0 >.
z E-"







= 4 0
0 a 0

S0 4 o1 a t (0 3

14 P > >U
0 04


-0 &
l- 03

0e I
1 01 C -4
0 (01 0 414

.(4 4-
>i 0 i4 0 0
-4 3 0 '
0 >

)a l m 4C: 0 1

0' > = >
06 A >C

a -4
> >

-44 4

> C


J 0

o X


-0 .
0 t

'a -4

a0 Mo C
U 0 H )

------- E-'14



4 1l

0 0
PI o







l ^






I -14

< 0.. a
Cl 1. 3 1C
& U0
3 Li4






U (0
0-4 Hr

f l



o 0)4



r --4 0)

O -4 40

H 4

50 "-4
O 0'w

0 4-.
w C


41 44 04

O 3
0 0
0 fl '


c0 4)

'C a

Q -4
u <


> 0


H( 144





from both experiments can also be used to support a model

with six membrane spanning regions. The ambiguity of the

alkaline phosphatase technique can explain the differing

interpretations of fusion protein data. Evidence for the

amino-terminus being located on the periplasmic facing of

the membrane was obtained via fj-galactosidase fusions with

the a subunit (Lewis & Simoni, 1992).

Significance of the a Subunit

FIFo ATP synthase is of biological importance to the

bioenergetics of all aerobic organisms. The a subunit is

essential for the function of FIFo ATP synthase. The

importance of FIFO ATP synthase and the a subunit were

demonstrated by the NARP mutation in human mitochondrial

genetic diseases. High levels of the NARP mutation in the

a-like subunit in human mitochondria correlates with a

clinical phenotype of neurological atrophy. The phenotype

is consistent with a reduction in cellular energy levels

resulting in the death of nerve cells.

The a subunit is also essential to the mechanism of

proton translocation. The mechanism of proton transport is

probably similar to the mechanism of transporting other

ions. This is supported by the FIFo ATP synthase of P.

modestum, which transports sodium ions as well as protons.

Ion transport is a vital part of all biological systems, and

very little is known about how it is accomplished and


The a subunit may also play a role in coupling proton

translocation with ATP catalysis. How energy is converted

from an electrochemical gradient into the chemical bond

energy of ATP is one of the primary unanswered question

concerning the chemiosmotic hypothesis.

The conservation of all FIFo ATP synthases at the

functional and structural level indicate that the E. coli

FIFo ATP synthase can be used as a model for investigating

the function of the a subunit. This work encompasses the

investigation of amino acids in the a subunit involved in

proton translocation (Chapter 4). The biochemical effects

of the ATPase 6 gene mutations, that result in the human

mitochondrial genetic disease NARP, were also determined by

modeling the mutations in the E. coli a subunit (Chapter 5).

Modeling of sequence variations that occur in the a-like

subunits of chloroplast and alkaliphilic bacteria revealed a

role for the a subunit in coupling proton translocation with

ATP catalysis (Chapter 6).


Molecular Techniques


T-4 DNA ligase, T-4 polynucleotide kinase, Taq

polymerase, and restriction endonucleases were supplied by

Bethesda Research Laboratories (Bethesda, MD) and New

England Biolabs (Beverly MA). Taq-Traq sequencing system

was the product of Promega (Madison, WI). AmpliTaq Cycle

Sequencing Kit was purchased from Perkin Elmer Cetus

(Norwalk, CT). Radionucleotides were purchased from

Amersham Corp. (Arlington Heights, IL) or ICN Biomedicals,

Inc. (Costa Mesa, CA). Lysozyme was supplied by Sigma

Chemical Co. (St. Louis, MO). Difco Laboratories (Detroit,

MI) was the source of bacterial growth media. Proteinase K

was purchased from Amresco (Solon, OH). All other reagents

and chemicals were obtained from Sigma or Fisher Scientific

(Orlando, FL). Oligonucleotides were synthesized in the

core facility of the Interdisciplinary Center for

Biotechnology Research at the University of Florida.

Organisms. Media and Growth Conditions

Bacterial strains and plasmids are listed in Table 2-1.

Luria broth containing 0.2% w/v glucose (LBG) was used as

the primary medium for DNA purification unless otherwise

noted. The concentration of chloramphenicol in liquid

medium was 15-20 Ag/ml and 30 Ag/ml in solid medium. Liquid

cultures were aerated by continuous mixing on an orbital

shaker or in a roller drum. Incubations were performed at


DNA Preparation

Large quantities of plasmid DNA were purified using a

CsC1 step gradient (Garger et al., 1983) and smaller

quantities were prepared by the rapid screen method

(Birnboim & Doly, 1979). Plasmid DNA for sequencing was

prepared by a modification of the rapid screen method.

Cells were grown in 10 ml Terrific Broth [1.2% bacto-

tryptone (w/v), 2.4% bacto-yeast extract (w/v), 0.04%

glycerol (v/v), 0.17 M KH2PO4 and 0.072 M K2HPO4] overnight

(Tartof & Hobbs, 1987). Cells were lysed and plasmid DNA

was recovered essentially as described by Birnboim and Doly

(1979). The plasmid DNA was treated with 10 units of RNAse-

It (Stratagene, La Jolla, CA) for 30 min, and then treated

with Proteinase K overnight in the presence of 25 mM EDTA.

The solution was extracted twice with phenol:chloroform

(1:1) and twice with chloroform. The DNA was precipitated

Oby the addition of 0.5 volume of 7.5 M ammonium acetate and

Table 2-1
List of Strains and Plasmids

Strain/ Source/
plasmid Genotype/description Reference







bglR thi-1 rel-1 HfrPO1
[a, c, b, 6, a, 7, 3, E]j
a, (uncB2000), c,.
in 1100 background
F AlacU169 araD139
thiA rpsL relA
ad (uncB2000)
in 1100 background
aV2AP24.WW24 (uncB205)
recA56 srl::TnlO
in 1100 background

Ap' a
Cmr aL07-WMJ09
Apr Cmr Tetr
Cam ori"
Cm' ad (uncB2000)
Cma a
Cma ad, c, b, 6, a, 7, 8, E
Cmr ad, c, b, 6-

Cmr a, NARP mutations
pBDC26 based plasmids
Cmr ae2 mutations
pPH12 based plasmids
Cmr af mutations
pPH12 based plasmids
Cm' a 9 mutations
pPH12 based plasmids
Cm' an mutations
pPH12 based plasmids
Cmr aG1 mutations
pBDC26 based plasmids
Cm' a,, mutations
pPH12 based plasmids

Humbert et al., 1983

Cain & Simoni, 1986

Kumamoto & Simoni,
Chapter 3

Humbert et al., 1983

Cain & Simoni, 1986
Cain & Simoni, 1989
Hamilton et al., 1989
Chapter 2
Chapter 2
Chapter 2
Chapter 2

Chapter 5

Chapter 4

Chapter 4

Chapter 4

Chapter 4

Chapter 6

Chapter 6

"New England Biolabs (Beverly, MA)


3 volumes of ethanol and recovered by centrifugation at 12,000

x g for 20 min. The pellet was rinsed with 70% ethanol, dried

under a vacuum and suspended in TE buffer [10 mM Tris-HC1, 1

mM EDTA, pH 8.0].

Genomic DNA was isolated by the method of Porter et al.

(1985). A 5 ml culture of the desired strain was harvested,

then suspended in 0.4 ml of GTE Buffer [50 mM glucose, 25 mM

Tris 10 mM EDTA, pH 8] containing fresh lysozyme [2 mg/ml] and

incubated at 4C for 30 minutes. Samples were heated to 700C

for 3 min, and lysed by adding 0.8 ml 1% SDS with gentle

mixing. Samples were cooled to 4C, then extracted by gentle

mixing with 1.2 ml phenol. Genomic DNA was collected by

spooling DNA at the interface of 2 volumes of ethanol layered

over the aqueous sample. The spooled DNA was rinsed with 80%

ethanol and allowed to dry for a few minutes before suspension

in TE buffer. An aliquot of the preparation was extracted

with phenol, and then extracted with ethyl ether to remove the

phenol. The DNA was precipitated in 2 volumes of ethanol, and

dissolved in TE buffer. Aliquots of the DNA were used for

either Southern analysis or for genomic sequencing.

PCR Amplification of DNA

Polymerase Chain Reaction (PCR) was used to amplify

genomic DNA isolated from strain RH305 (a,,,a) for sequencing

of the uncB205 mutation. The uncB gene was amplified using

oligonucleotide primers (GGTGCTGGTGGTTCAGATACTGGCAC,



conditions (Saiki, 1990). The reaction mixture consisted of:

100 AM dNTP's, 20 pmol of each primer, 1 ng of genomic DNA

template, 2.5 U of Taq polymerase, in Tag polymerase Buffer

[10 mM Tris-HCl, 50 mM KC1, 1.5 mM MgCl2, 0.001% (w/v)

gelatin, pH8.8]. Genomic DNA was amplified with 30 cycles of

the following temperature steps: 950C for 30 seconds, 65C

for 30 seconds, 720C for 1 minute. PCR products were

extracted with chloroform to remove mineral oil and

precipitated with 2 volumes of ethanol. The nucleotide

sequence of the PCR product was determined using cycle

sequencing according to the protocol from AmpliTaq Cycle

Sequencing Kit (Perkin Elmer Cetus, Norwalk, CT), as described


Restriction Diaest

Restriction endonuclease digestions were carried out

under conditions specified by the vendor. Conditions for

digestions involving more than one restriction endonuclease

were chosen to prevent any non-specific activities.

Restriction digested DNA used for ligations, mutagenesis

or nucleotide sequencing was handled as follows. Digested DNA

was diluted to 50 Al or 100 Al with TE buffer and extracted

with 1 volume of phenol/chloroform, followed by two

extractions with 1 volume chloroform. The DNA was

precipitated at room temperature for an hour in 2.5 M ammonium

acetate and 67% ethanol (by addition of 0.5 volumes of 7.5 M

ammonium acetate and 3.0 volumes of ethanol). The


precipitated DNA was pelleted (12,000 x g, 30 min) and dried

under vacuum to remove excess ammonium acetate. When small

amounts of DNA were digested, 20 Ag of tRNA was added to the

precipitation step to increase recovery.

DNA digested for analysis via gel electrophoresis was

handled differently. Upon completion of restriction

digestion, loading dye [1% bromophenol blue, 1% xylene cyanol,

50% glycerol] was added at 1/10 volume of reaction. The

sample was loaded into a well in an agarose horizontal gel

(usually 0.8% agarose) then subjected to electrophoresis. DNA

was visualized following staining with ethidium bromide via UV

light. Samples were stored at -200C upon completion of the

digest if not used immediately in electrophoresis experiments

(Sambrook et al., 1989).

Southern Analysis

DNA fragments size fractionated in a 0.6% agarose gel

were transferred through capillary action in 10X SSC onto Gene

Screen hybridization transfer membrane (NEN; Boston, MA). DNA

was cross-linked to Gene Screen membrane by UV light.

Prehybridization was in hybridization buffer [7% SDS, 1% BSA,

1 mM EDTA, 0.5 M NaH2PO4, pH 7.2] for 10 minutes at 65 C.

Hybridization was initiated by addition of radiolabled probe

in hybridization buffer, and incubation was allowed to

continue over night. The probe was made from the 1.1 kilobase

Sall fragment generated from the digestion of plasmid pEMS54

(unc operon), and radiolabeled according to the random primers


kit (BRL) directions. The hybridization membrane was washed

with buffer [ImM EDTA, 40 mM NaH2PO4, pH 7.2] until non-

specific bound radiolabled probe was removed. This was

accomplished by monitoring radiation released into the wash

buffer. The signal was visualized by exposing Kodak XAR 5

film to the hybridization membrane overnight


Liqations and Transformations

Ligations were preformed by three different protocols,

depending on the source of the ligase and the type of

ligation. For 'sticky-end' ligations using NEB T4 ligase, 1

jg of DNA was incubated with 0.2 pl of ligase (80 U) in 50 Al

of ligation buffer [40 mM Tris-HC1, 10 mM MgC12, 10 mM DTT,

0.5 mM ATP, pH 7.5]. The reaction was incubated over night at

160C to facilitate annealing of restriction site overhangs.

Insert to vector ratios were 1, 3 and 10 fold molar excess.

For 'blunt-end' ligations 4.5 il of NEB T4 ligase (1800 U) in

ligation buffer was incubated at room temperature to increase

T4 ligase efficiency. When BRL T4 ligase was used for sticky-

end ligations a different protocol was used. The DNA

concentration was dropped to 0.1 pg, and only 0.1 U of BRL

ligase was used in a 20 p1 reaction volume with BRL ligase

buffer [50 mM Tris-HCl, 10 mM MgCl,, 1 mM DTT, 1 mM ATP, 5%

(w/v) polyethylene glycol-8000, pH 7.6]. The incubation was

one hour at room temperature. The insert to vector ratios


were 1, 3 and 5 fold molar excess. The ligation reaction was

usually diluted 5 fold with TE before use.

Competent cells were prepared by harvesting a mid-

logarithmic phase culture by centrifugation (8000 x g, 5 min,

40C) and treating the E. coli with cold (40C) 50 mM CaC12.

Treatment was with 0.5 volume of cold 50 mM CaCl2 for 45

minutes, followed by storage in 0.1 volume of cold 50 mM CaC12

for 4 to 24 hours before use.

Transformations were performed as follows. Competent

cells were added to DNA (1/3 to 1/2 of a ligation reaction

mixture, 1 Al of a plasmid preparation) and incubated at 40C

for a half an hour. The cells were then heat shocked at 420C

for two minutes. LBG medium was added to the cells and they

were allowed to grow aerobically by incubation of the cultures

for 1 to 2 hours at 370C. The cells were then harvested by

centrifugation and spread on LBG solid medium containing the

appropriate antibiotic.

Strain MC4100 (Table 2-1) was used for transforming

ligation reactions, and conditions were optimized for its

efficiency. Transformations into other strains for specific

applications used DNA from plasmid preparations.

Plasmid Construction

Plasmid DPH1. Plasmid pPH1 (Amp', Cm'; 5.1 kb) was

constructed from plasmid pBR325 (Tet', Amp', Cm') (Table 2-1),

by removing the 375 bp BclI/BamHI fragment. This was

accomplished by digestion of plasmid pBR325 with Bcll and


BamHI, and circularizing by ligation. The deletion of the

BclI/BamHI fragment was confirmed by restriction site mapping

and the loss of tetracycline antibiotic resistance.

Plasmid DPH4. Plasmid pPH4 (a, Cmf; 5.4 kb) was

constructed by lighting the 1070 bp AatII/AseI fragment

containing the uncB (a) gene from pBDC1 (Cain & Simoni, 1986),

into the 4.36 kb AseI/AatII fragment of pPH1 (Amp', Cm'). The

construction of the plasmid was confirmed by restriction site

mapping, loss of ampicillin resistance, and the ability to

complement the uncB205 mutation in strain RH305.

Plasmid DPH5. Plasmid pPH5 (a, Cm'; 4.6 kb) was

constructed by digesting plasmid pPH4 (a) with Narl

restriction endonuclease, and then circularizing the 4638 bp

fragment of the plasmid under sticky-end ligation conditions.

The 4 Narl restriction sites present in plasmid pPH4 (a) were

reduced to one in plasmid pPH5 (a). The deletion of the DNA

between the Narl sites was confirmed by restriction site


Plasmid PPH6. Plasmid pPH6 (a, Cm'; 4.6 kb) was

constructed by destroying the Clal and HindIII restrictions

sites in plasmid pPH4 (a, Cm'). This was accomplished by

digesting pPH4 with ClaI and HindIII, then incubating with DNA

polymerase I (Klenow fragment) and dNTP's to fill in the 3'

recessed ends left by the restriction enzymes. The DNA was

circularized by lighting under blunt-end conditions (see

above). The construction was confirmed by restriction site


mapping, and the plasmid was shown to complement the uncB205

mutation in strain RH305.

Plasmid DPH7. Plasmid pPH7 (a, Cm'; 4.6 kb) was

constructed by removing the Narl restriction sites from pPH4

(a). This was accomplished by digestion with Narl followed by

filling the recessed ends with DNA polymerase I (Klenow

fragment) and dNTP's. The DNA was circularized under blunt-

end ligation conditions. The removal of the NarI restriction

sites from the plasmid also decreased the number Ahall

restriction sites, facilitating the construction of plasmid


Plasmid DPH9. Plasmid pPH9 (a-, Cm'; 4.4 kb) was

constructed to remove the first Ahall restriction site in the

uncB (a) gene. Plasmid pPH7 (a) was digested to completion

with restriction enzyme AatII and partially with restriction

enzyme Ahall. The larger fragment was isolated, containing

the cleavage of only the first AhaII restriction site in uncB

(a) gene and the AatII restriction site. The fragment was

treated with DNA polymerase I (Klenow fragment) and dNTP's to

fill-in the 3' recessed ends. The plasmid was circularized

under blunt-end ligation conditions. Plasmid pPH9 (a-) now

contained only one AhaII restriction site in the carboxyl

terminal region of an incomplete uncB (a) gene.

Plasmid DPH10. Plasmid pPH10 (a-, Cm'; 4.4 kb) was

constructed by cassette mutagenesis (see below) as follows.

Plasmid pPH9 (a-) was digested with the restriction enzymes


BclI and Ahall. A synthetic oligonucleotide cassette

containing silent mutations that generated Clal and HindIII

restriction sites was ligated to the --kb BclI/AhaII fragment

of plasmid pPH9 (a-) (Figure 2-1). Plasmid pPH10 (a-) carried

an incomplete uncB (a) gene containing unique restriction

sites that could be used for site-directed cassette


Plasmid DPH11. Plasmid pPH11 (a, Cm'; 4.0 kb) was made

by the digestion of plasmid pPH5 (a) with BamHI, and then the

4.0 kb fragment was circularized by sticky-end ligation.

Construction of plasmid pPH11 (a,) was confirmed by

restriction site mapping and loss of complementation of

strains PH105 and RH305 on succinate minimal medium.

Plasmid DPH12. Plasmid pPH12 (a; Cmr; 4.6 kb) was made

to restore the disrupted uncB (a) gene of plasmid pPH10 (a')

using the uncB (a) gene from plasmid pPH6 (a). Both plasmids

were digested with the restriction enzymes PstI and Aval and

the appropriate DNA fragment isolated. The 2.2 kb PstI/AvaI

fragment of plasmid pPH6, containing the amino-terminal

section of the uncB (a) gene and the chloramphenicol

resistance gene, was ligated with the 2.4 bp PstI/Aval

fragment of plasmid pPH 10, containing the plasmid origin of

replication and the carboxyl-terminal section of the uncB (a)

gene. Plasmid pPH12 contained the complete uncB (a) gene with

the silent mutations for the unique restriction sites Clal and

HindIII (Figure 2-1). The construction was confirmed by

N E-4
P 0 H

0U 0

0 U

0 <

oE- E-
00 u
4 0


0 0

U 0

u 0

0 0U

m 4
F 4


(^ U U U
PQ ^^83^


O1) 0

* .. 0

0 Q4$*
0 0
* -0 C

OV t 4 0
I At V

0 0- 0
W) 44
."0) u

4) UONi 4)
a a,

0 u 0
) a ) ic

V.1 .- 04
S10 04
0 uC

-4 4 4 *H .
*4 V4 010
tQC 0

t u o t
"I ON 4)
) 4 1a) *. U

C4 4r ta 0M
m 0, 0
3 *< c-

z 0 7 0
0 0W
A-4) cd 0

V4 .4 -Q4
<0 (0 .I )

- 44 4) 1 Q)
0 C *S

V0 Q) 0< 40
U)e 10 (
1O U ^.C
a a0 0 0
> C


0'Q 0 3 4
*1O *d
3 D0'
*e 0
4C 2.4 #

restriction site mapping and complementation of the a

subunit deficient strains PH105 and RH305.

Plasmid pPH13. Plasmid pPH13 (a,,c,b,6,a,7y,1,e, C';

13.6 kb) was constructed by lighting the 2.1 kb BamHI/XbaI

fragment from pAP55 (a,c,,b,6,a,7,j,e), unc promoter, uncI

and part of the uncB (a) gene, with the 11.5 kb BamHI/Xbal

fragment from pEMS54 (a,c,b,6,a,7,3,e), containing part of

uncB (a), uncE (c), uncF (b), uncH (6), uncA (a), uncG (7),

uncD (f), uncC (e), the Cm resistance gene, and origin of

replication for plasmid pACYC184. The resulting plasmid

contained the complete unc operon minus the 617 bp BamHI

fragment from uncB (a), which was confirmed by restriction

site mapping.

Plasmid pPH15. Plasmid pPH15 (a c, ,b, 6-, Cmf; 6.9

kb) was constructed by ligation of the 2.2 kb HindIII/ClaI

fragment from plasmid pPH13 (aM,c,b,6,a,y,3,E) into plasmid

pMAK705 digested with HindIII and Clal. Plasmid pMAK705 was

kindly provided by Sidney R. Kushner (University of


Cassette Site-Directed Mutagenesis

Cassette site-directed mutagenesis requires that

restriction endonuclease sites bracket the region of

interest. These restriction sites should be unique in the

mutagenesis vector. The segment of DNA between the

restriction sites constitutes the 'cassette', which can now

be replaced following restriction enzyme digestion, and

ligation with a synthetic double-stranded oligonucleotide

cassette. The double stranded oligonucleotide cassettes

were made by annealing complementary synthetic

oligonucleotides that encoded the desired mutation(s) and

generated sticky-ends compatible with the restriction

fragment ends. The cassette allows for multiple changes,

including incorporation of 'doped' nucleotides at a given

positionss. Use of a 'doped' cassette results in a mixture

of oligonucleotides which are identical except for the

positions reflecting ambiguous synthesis. From this type of

mutagenesis cassette many changes at a given codon can be

constructed at once, resulting in an array of amino acid

substitutions. Each clone from an experiment of this type

was sequenced to determine which mutation had been


Several cassettes and cassette mutagenesis plasmids

were used in this thesis and are detailed in each chapter.

Figure 2-1 shows the cassette used to construct plasmid

pPH12 resulting in the HindIII and Clal sites incorporated

into the carboxyl terminal region of the uncB (a) gene.

PCR Mutagenesis

PCR assisted mutagenesis was used to construct mutants

in Chapter 6. The basic protocol was as follows. A sense

primer was designed upstream of the restriction site PstI in

the uncB (a) gene (Figure 2-1). An antisense mutagenic

primer was designed that contained the am2,e mutation. This

primer also encompassed the BclI restriction site in the

uncB (a) gene (see Figure 2-1). The BclI restriction site

was 3' with respect to the a.us4 mutation, allowing the PCR

product containing the mutation to be ligated into the uncB

(a) gene. Standard PCR conditions were used (see above) to

amplify DNA from the wild-type plasmid pPH12 (a) and the

mutagenic plasmids pUNCB5.40 (a1DD) and pUNCB5.41 (aG1K)

separately (Figure 2-1; Chapter 6, Table 6-1). The PCR

products were ligated into pPH12 (a) using the restriction

enzymes PstI and Bcll (by the protocols detailed above), and

the nucleotide sequence was confirmed across the whole

cloned region using the the AmpliTaq system (see below).

Sequencing Reactions

Taa-Traq. DNA sequencing was carried out using the

Taq-Traq sequencing system for the mutations in chapter 4.

Briefly, 2-4 ig of plasmid DNA in a total volume of 18 AL

was denatured by alkali treatment with 2 gL of 2 N NaOH for

5 min at room temperature. Base was neutralized by mixing

with 2 AL of 3 M ammonium acetate (pH 4.8). DNA was

precipitated in 75 AL of ethanol at -70C for 10 min, and

microfuged (12,000 xg) for 15-30 min. The DNA pellet was

rinsed with 200-500 pL of cold 75% ethanol and dried under

vacuum for 5-10 min. DNA pellet was suspend in 10-16 AL of

water, stored on ice and used as quickly as possible to

minimize the renaturing of DNA strands. The denatured DNA

template was added to 2 pmol of sequencing primer, 5 AL of


TAQ polymerase 5X reaction buffer, and 2 tL of the extension

mix to give a volume of 25 AL. The primer was annealed to

the template for 10 min at 500C. 2 AL of radioactivity

(dATPa"S: 1,000 Ci/mmol, 10 pCi/fL) and 1 AL of TAQ

polymerase (sequencing grade) were added, mixed, and

incubated at annealing temperature for 2 min. 6 AL of the

extension mixture was aliquotted into the four termination

tubes (labeled A C G T) containing the appropriate

terminating ddNTP, mixed and incubate at 700C for 5 min.

The reaction was stopped by adding 4 ML of the Stop

Solution. If not used immediately, samples were stored at -

20C and used within 2 weeks. Samples were heated to 950C

for 2 min before size fractionation by electrophoresis in a

6% acrylamide, urea denaturing, sequencing gel (Garoff &

Ansorge, 1981).

AmpliTaq. Cycle sequencing was done according to the

protocol from AmpliTaq Cycle Sequencing Kit (Perkin Elmer

Cetus, Norwalk, CT). Approximately 1 Mg of Plasmid DNA that

had been previously digested with PvuII (see above) in a

total volume of 15 Al was added to AmpliTaq buffer with 2

pmol of primer; end-labeled with ATP-y-32P and kinase, then

purified over a Nensorb-20 column (New England Nuclear;

Wilmington, DE). The mixture was split into 4 tubes

containing the appropriate ddNTP and amplified with 20

cycles of the following temperature steps: 15 sec at 95C,

1 min at 500C, 1 min at 720C; followed by 10 cycles of the


following temperature steps: 15 sec at 95C, 1 min at 720C.

Reactions were stored at 40C until stop solution was added.

If not used immediately, samples were stored at -20C and

used within 5 days. Samples were heated to 95C for 2 min

before size fractionation by electrophoresis in a 6%

acrylimide, urea denaturing, sequencing gel (Garoff &

Ansorge, 1981).

Membrane Preparations


Difco Laboratories (Detroit, MI) was the source of

bacterial growth media. Deoxyribonuclease I was purchased

from Sigma Chemical Co. (St. Louis, MO). All other reagents

and chemicals were obtained from Sigma or Fisher Scientific

(Orlando, FL).

Cell Fractionation

Preparation of subcellular fractions was as described

previously (Klionsky et al., 1983). E. coli were inoculated

into LBG (500 ml) and grown to a cell density of

approximately 150 Klett units. The cells were harvested by

centrifugation (5000 x g, 5 min), washed with TM buffer (50

mM Tris-HCl, 20 mM MgSO4, pH 7.5) and suspended in 7 ml of

TM buffer containing DNase I (10 Ag/ml). Cells were

disrupted at 14000 psi in a French pressure cell. Debris

and unbroken cells were removed by two successive

centrifugation (8000 x g, 5 min). The particulate membrane

fraction and soluble cytoplasmic fractions were separated by

high speed centrifugation (150,000 x g, 1.5 h).

Membranes used for proton conduction studies were

prepared by suspending the membrane pellet in 1.5 ml of TM

buffer by use of ground glass homogenizers. The suspended

membranes were then diluted to a total volume of 7 ml with

TM buffer then pelleted again by high speed centrifugation

(150,000 x g, 1 hr). Pellets were suspended as above in

1.2 ml of TM buffer. Samples were centrifuged (12,000 x g,

5 min) to remove any remaining cell debris and the supernate

was transferred to a new tube for use in fluorescence


Membranes used for Fi stripping experiments were

suspended in SB buffer [1mM Tris-HC1, 0.5 mM EDTA, 2.5 mM 2-

mercaptoethanol, 10% (v/v) glycerol, pH 8.0]. The stripped

membranes prepared for analysis in chapter 4 were diluted to

a total volume of 7 ml with SB buffer then pelleted again by

high speed centrifugation (150,000 x g, 1 hr). The stripped

membranes prepared for analysis in chapter 5 and 6 were

diluted to a total volume of 7 ml with SB buffer and gently

agitated for 1 hr before being pelleted again by high speed

centrifugation (150,000 x g, 1 hr). All stripped membranes

preparations were suspended and diluted in SB buffer as

above and gently agitated overnight. Stripped membranes

were recovered by centrifugation (150,000 x g, 1 h) and then

suspended in 1.2 ml of TM buffer. The membranes were

centrifuged (12,000 x g, 5 min) to remove any remaining cell


Three separate procedures were used to prepare

membranes used for measuring F, ATPase hydrolysis activity.

The method used in chapter 3 and 4 involved simply

suspending the membrane pellet from subcellular

fractionation in 1.2 ml of TM buffer. Membranes for ATPase

hydrolysis activity from the first part of chapter 5,

considering the awR mutation, were prepared by an improved

method. The membrane pellets were suspended in PSB buffer

(5mM Tris-HC1, 0.5mM EDTA, 15% (v/v) glycerol, 2.5 mM 2-

mercaptoethanol, 6 mM p-aminobenzamidine, pH 7.5) diluted

and gently agitated for one hour before being pelleted again

by high speed centrifugation (150,000 x g, 1 hr). The

membrane pellets were once again suspended in PSB buffer,

diluted up to 7 ml, and gently agitated overnight prior to

pelleting by high speed centrifugation (150,000 x g, 1 hr).

The last step was to suspend the membrane pellet in 1.2 ml

of TM buffer. The membranes prepared for the second part of

chapter 5, the a ep mutation, and the membranes for Chapter

6 were prepared in the same manner with two modifications to

improve the dissociation of uncoupled F, ATPase. First, the

membrane 'washing' step, suspending the membrane pellets in

PSB buffer and gently agitating for one hour prior to

pelleting, was performed twice. Second, after the final

suspension in 1.2 ml of TM buffer, the membranes were

centrifuged (12,000 x g, 5 min) to remove any remaining cell

debris. All steps performed at 4C.

Protein Assay

Protein concentrations were measured using the modified

Lowry procedure of Markwell et al. (1978). An aliquot of

membrane vesicles was diluted 1:10 or 1:5 with TM buffer.

Aliquots of 5, 15 and 25 ML were taken from the dilution for

assaying. A standard curve of bovine serum albumin (BSA)

was used with a concentration range of 5 to 100 Ag of

protein. Duplicate samples were incubated in 3 ml of

alkaline copper [100 parts reagent A (2.0% Na2CO3, 0.4% NaOH,

0.16% sodium potassium tartrate, 1% SDS) to one part reagent

B (4% CuSO4)] for 15 to 60 minutes at room temperature.

Color was developed by adding 0.3 ml of Folin-Ciocalteu

phenol reagent (diluted 1:1 with distilled water immediately

prior to addition) and incubating 45 min at room

temperature. Absorbance at 660 nm was read in a Pharmacia

spectrophotometer, and the duplicate samples averaged.

Protein concentrations were determined from the linear range

of the standard curve by using a linear regression program.

Membrane protein concentrations were usually 15 to 30 mg/ml

for unstripped membranes and 2 to 5 mg/ml for stripped


Storage Conditions

Membrane preparations were stored at 40C and most

biochemical assays were preformed within 24 hours of the

cells being lysed. If the membranes were to be stored for a

period longer than 24 hours, the preparations were quick

frozen in liquid nitrogen and kept at 80C. Frozen

membranes were then thawed in a 370C water bath before use.



Pyruvate kinase and lactate dehydrogenase were purchased

from Sigma Chemical Co. (St. Louis, MO). 9-Amino-6-chloro-

2-methoxyacridine (ACMA) was purchased from Molecular Probes

(Eugene, OR). ECL Western blotting detection system, and

Anti-rabbit Ig linked to horseradish peroxidase were from

Amersham corp. (Arlington Heights, IL). All other reagents

and chemicals were obtained from Sigma or Fisher Scientific

(Orlando, FL).

Growth Characteristics

Minimal media employed for growth studies consisted of

minimal-A salts (Miller, 1972) supplemented with either

glucose or succinate (0.2% w/v) as specified in the text.

Growth on succinate was determined by visual inspection

of colony size after 2 4 days growth at either 37C or

420C. The scale by which colony size was rated is wild-type

or near wild-type colonies (+++), clearly smaller colonies

with substantial growth (+), minute colonies that were

barely visible but present (), and no visible colonies (-).

Plasmids carrying a subunit site-directed mutations were

transformed into E. coli a subunit defective strains PH105

and RH305. Mutant strains were grown on succinate media

multiple times and plasmids carrying mutations were

transformed into each strain at least twice. Growth on

succinate characteristics were used to check for reversions

to wild-type phenotypes (especially in strain RH305 for

reasons detailed in chapter 3).

Growth yields of strains were determined

turbidimetrically by a Klett-Summerson colorimeter, and

yields were reported in Klett units. Minimal A medium

(Miller, 1972) supplemented with 5mM glucose was used unless

otherwise indicated. The cultures were aerated by

continuous mixing on an orbital shaker at 370C.

Anaerobic Growth

Anaerobic growth conditions were sustained with the

GasPak H2 + CO2 generating system in an anaerobic jar.

Glucose minimal medium agar plates were incubated under

vacuum, followed by preincubation under anaerobic

conditions. Colonies were streaked on to the plates. The

plates were then placed into the anaerobic jar, in which the

02 content had been reduced with CO, subliming from crushed

dry ice. Anaerobic conditions were obtained (GasPak 0O

indicator strip) before the minimal medium plates were

placed in a 370C incubator. Strains were incubated for 48

hours before colony size was scored.

ATP-Driven Proton Translocation

Chapter 4 and 5. Membrane energization was assayed by

fluorescence quenching of 9-amino-6-chloro-2-methoxyacridine

(ACMA) (Aris et al., 1985). Membrane vesicle concentration

was 250 Ag protein/ml buffer [50 mM MOPS and 10 mM MgCl2, pH

7.3] in a volume of 3 ml buffer. ACMA was added to final

concentration of 1 AM, and a Perkin-Elmer Model LS3B

Fluorescence spectrometer was used to measure the change in

ACMA fluorescence over time. Samples were first treated

with NADH (0.5 mM) resulting in fluorescence quenching due

to an acidification of the vesicles driven by the electron

transport chain. NADH driven quenching was used to test the

integrity of the membrane preparation (usually not shown in

figure). Fluorescence was recovered by poisoning the

electron transport chain with KCN (0.5 mM). After the

fluorescence had recovered ATP (0.4 mM) was added to assay

ATP-driven vesicle acidification. Fluorescence was once

again recovered using nigericin (0.05 AM) to collapse the

proton gradient. Treatment of fractionated membranes with

50 AM DCCD for 15 minutes at 370C was used for inhibition of

FFo ATP synthase proton translocation.

Chapters 5 and 6. Membrane energization was done

essentially as described above with the following

modifications. Membrane concentration was 150 Ag of

membrane protein/ml buffer [50 mM Tris and 10 mM MgC12, pH

7.5]. Fluorescence quenching was driven by addition of 500

MM ATP, and addition of 100 iM DCCD was used to recover the

fluorescence by inhibiting FIFo ATP synthase proton

translocation. Analysis of pH conditions and ATP

concentrations affect on proton pumping (Chapter 6) were

done essentially the same, but membrane integrity was not

confirmed with NADH for each ATP concentration. However,

each membrane preparation was tested for NADH-driven

fluorescence quenching in a separate trace. Additionally,

the buffer used was pH 7.2 instead of pH 7.5. Both these

changes were made to increase the apparent rate of the

relative fluorescence change (see Chapter 6).

Passive Proton Conductance

Membrane vesicles stripped of F, (described above) were

acidified with NADH (0.5 mM) to test passive proton

conductance in mutant a subunit Fo sectors. Wild-type Fo

sectors exhibit passive proton conductance, which decreases

NADH driven ACMA fluorescence quenching by collapsing the

generated proton gradient. Stripped membrane vesicles were

analyzed under the same conditions as indicated for the

intact membrane vesicles of the same mutant strains.

ATPase Activity

Inorganic phosphate assay. ATPase activity was

measured by the release of inorganic phosphate from the

hydrolysis of ATP. Inorganic phosphate was assayed by the

acid-acetone-molybdate method of Heininen and Lahti (1981).

Inorganic phosphate complexes with the molybdate in an acid

form that absorbs light at 335 nm in organic solvent. The

molybdate-phosphate complex was stabilized by the addition

of citric acid, and will degrade very slowly. The citric

acid also complexes with all the free molybdate, preventing

ATP hydrolysis by the acid in the mixture from interfering

with the assay of enzymatic activity. Moreover, the F,-

ATPase activity was abolished by the acidic mixture, so

addition of the enzyme assay buffer to the acid-acetone-

molybdate solution stopped the enzyme reaction.

Phosphate contamination was minimized by use of new or

acid washed glassware to prevent interference with the

assay. ATPase activity was measured under the following

conditions: 70 Ig membrane vesicle protein, in 3.5 ml of

buffer [50 mM Tris-HCl, 1 mM MgCl2, pH 9.1] at 37C. At

time zero 3 mM ATP was added and aliquots of the enzyme

solution taken at 1, 2, 5, 10, 15, 20 minutes and were added

to the acid-acetone-molybdate solution, effectively stopping

enzymatic ATP hydrolysis. A time zero aliquot was used to

blank out background phosphate contamination. Inorganic

phosphate concentrations at each of the time points were

determined by comparison of absorbance to the absorbance of

a phosphate standard curve. Specific activity was expressed

as Amol of ATP hydrolyzed/min/mg membrane protein.

Coupled assay. ATP hydrolysis was also measured by a

coupled assay to regenerate ATP in the system. Membranes

(15 or 30 ig protein) were added to the assay buffer (2mM

MgCl2, 25 mM KC1, 5mM KCN, 25 mM Tris-HC1, 5 units/ml

pyruvate kinase, 5 units/ml lactate dehydrogenase, 2.5 mM

phosphoenol pyruvate, and 0.5mM NADH, pH 8.0) and monitored

by absorbance at 350 nm for loss of NADH upon addition of

ATP-MgC1l (2mM). One mol of ATP was assumed to be

hydrolysis for each mol of NADH oxidized. Specific activity

was expressed as pmol of ATP hydrolyzed/min/mg of membrane

protein. Treatment of fractionated membranes with 50AM DCCD

for 1 hour at room temperature was used to inhibit FIFo

ATPase activity. DCCD insensitive activity was considered

to be non-specific binding of FI-ATPase to the membrane.

Western analysis. Twenty gg of membrane vesicle

protein samples were size fractionated by SDS-PAGE using a

10% tris-tricine acrylamide gel (SchAgger & Jagow, 1987) in

a Bio-Rad Mini Protean II Electrophoresis System. Samples

were subjected to electrophoresis till the bromophenol blue

tracking dye reached the bottom of the gel. Protein was

transferred from the gel to nitrocellulose in a tris-glycine

transfer buffer [20 mM Tris-HCl, 150 mM glycine, 20%

methanol, pH8.3] with a Bio-Rad Mini Trans-Blot

Electrophoretic system at constant amperage of 100 mAmps for

16 hrs at 40C (Towbin et al., 1979). Nitrocellulose blots

were stained with Fast Green and the sample lanes were cut

out and blocked in a buffer of 5% Carnation non-fat dried

milk in TBS [10 mM Tris-HC1, 150 mM NaCl, 0.02% sodium

azide, pH 7.2] for 2 hrs at room temperature (Tamarappoo et

al., 1992). Nitrocellulose blots were then incubated with

1/500 dilution of anti-Fi antibodies (Brusilow et al., 1981;

generously provided by R. D. Simoni) or a 1/5000 dilution of

anti-b antibodies (Deckers-Hebestreit et al., 1992;

McCormick et al., 1993; Stack & Cain, 1994; generously

provided by K. Altendorf) in blocking buffer for 1.5 hrs.

Blots were washed with TBS-Tween 20 [TBS + 0.3% Tween 20],

then incubated with 1/50,000 dilution of Anti-rabbit Ig

linked to horseradish peroxidase (Amersham) for 45 min in

blocking buffer. Blots were washed again with TBS-Tween 20,

then the ECL Western blotting detection system (Amersham)

was used to visualize the antibody blots. Autoradiographs

using Hyperfilm (Amersham) were exposed for 5 30 min.



The strain traditionally used for studies of the a

subunit, strain RH305, carries the uncB205 mutation in a

recA background (see Table 2-1). The chromosomal uncB205

mutation has been mapped by recombination to the 3' one-

third of the uncB (a) gene and can be suppressed by an amber

tRNA suppressor mutation (B. D. Cain personal

communication). The uncB205 mutation was sequenced from

strain RH305 by K. McCormick and determined to consist of

the following mutations: V239-A, P240-W and W241-end (TAG

stop codon). The ability of strain RH305 (a,~,) to undergo

reversion to a wild-type phenotype by an amber tRNA

suppressor or by reversion of the stop codon, made the

strain less than ideal for propagating uncB (a) mutation

plasmids through extended cell divisions. This was

especially true for uncB (a) mutations that impaired FIFo

ATP synthase function, because the mutant strain was placed

under selective pressure to produce reversion or amber

suppressor mutation. A wild-type or near wild-type

background of a subunits produced from reversion or tRNA

suppression in strain RH305 (aBa,,) would obscure the

phenotype of an a subunit encoded by a mutant plasmid that

effected activity. This in turn would have led to

inaccurate biochemical characterization of the mutants

effect on activity.

The construction of an uncB (a) gene deletion strain

would solve the problem of amber suppressors found in strain

RH305 (a.a). The uncB (a) gene deletion strain BC2000

reported by Cain and Simoni (1986), was found to have low

expression of F, subunits. This was thought to have been

due to a mutation in the Shine-Dalgarno sequence of the uncE

(c) gene (designated cm). The mutation originated in

plasmid pRPG54 which was the first unc operon expression

plasmid (Solomon & Brusilow, 1988). A derivative of plasmid

pRPG54 was used to construct strain BC2000 (a,cm,). The

mutation resulted in a polar phenotype reducing expression

of the rest of the unc operon genes. Plasmid pEMS54 was

constructed to correct the altered uncE (c) Shine-Dalgarno

sequence found in plasmid pRPG54 (Schaefer et al. 1989).

Plasmid pEMS54 was used as the source of the unc operon in

the construction of strain PH105 (ad) so that the reduced

expression of F, subunits could be avoided.

Construction of Strain PH105

The uncB (a) deletion strain PH105 was constructed via

homologous recombination between the chromosome of strain

1100 and plasmid pPH15 (a-,c,b,6-) essentially as described

by Hamilton et al. (1989). Plasmid pPH15 (a-,c,b,6-)

carries the uncB (a) gene deletion of 617 bp between the two

BamHI sites located in the uncB (a) gene, as well as, the

temperature sensitive origin of replication from plasmid

pMAK705. Strain 1100 carrying plasmid pPH15 (a-,c,b,6-) was

grown at the non-permissive temperature for plasmid

replication (440C) in the presence of the chloramphenicol

(Cm). The bacteria were passed three times to log phase in

order to obtain a co-integrate via homologous recombination

between the plasmid and the genome. The culture was then

returned to a permissive temperature for plasmid replication

(30C) in the presence of Cm. This step allowed a low

frequency secondary recombination event excising the plasmid

from the genome. Theoretically resolution of the co-

integrate should yield cells with either the normal uncB (a)

gene or the constructed deletion of the uncB (a) gene in the

genome. After three passes the culture was then returned to

the non-permissive temperature for plasmid replication

(440C). Antibiotic was withdrawn to remove the pressure to

retain the plasmid. The final step was designed to cure the

bacteria of the plasmid, since the plasmid is incapable of

undergoing replication at 440C. Finally, aliquots of the

culture were plated onto LBG media at different dilutions.

Of 41 colonies tested 3 did not grow on Cm media indicating

resolution of the co-integrate and curing of the plasmid.

One colony was also unable to support growth on minimal

medium supplemented with succinate, a non-fermentable carbon

source (Chapter 2). This colony was designated strain PH105

(a,) and was tested by complementation and Southern


Complementation Analysis of Strain PH105

Strain PH105 (a,) was tested for complementation by

plasmids containing different unc operon genes (Table 3-1).

Plasmid pAP55 (a, c, b, 6, a, 7, 8, e) carried all the genes

of the unc operon, and complemented strain PH105 (a,) for

growth on succinate minimal medium. Plasmid pPH13 (c, b, 6,

a, 7, 8, e) carried the unc operon minus the uncB (a) gene
and did not complemented strain PH105 (a,) for growth on

succinate minimal medium. Plasmid pPH12 (a) carried only

the uncB (a) gene, and complemented strain PH105 (aM) for

growth on succinate minimal medium. The results of the

complementation study indicated that a mutation in the uncB

(a) gene was present in strain PH105 (aM), and the mutation

was complemented by a wild-type uncB (a) gene carried on a

plasmid. None of the other unc operon genes carried on a

plasmid were sufficient for complementation of strain PH105

Table 3-1
Complementation Analysis

Strain/ Subunits Present Growth on
Plasmid strain/plasmid Succinate"

PH105 a,/none
PH105/pAP55 ad/a, c, b, 6, a, 7, 0, e +++
PH105/pPH13 au/c, b, 6, a, 7, 0, E
PH105/pPH12 a /a +++

RH305 a,,ay/none
RH305/pPH13 a, /c, b, 6, a, y, 3, e
RH305/pPH12 a, w/a +++

1100ABC unc,/none

Growth on minimal A medium supplemented with
succinate. Colonies were scored after 48 hours at 370C on
the as follows: wild-type growth (+++) or no growth (-).

(aM). Identical results were observed in the

complementation studies of the control uncB (a) gene

mutation strain RH305 (awaB) (Table 3-1).

Southern Analysis of Strain PH105

Confirmation that the uncB (a) deletion was

incorporated into strain PH105 (aM) was shown by Southern

analysis. Samples of wild-type strain 1100 and strain PH105

(ad) chromosomal DNA were digested with the restriction

endonuclease PstI and probed with the SalI fragment from the

unc operon (Figure 3-1). The Southern analysis showed the

predicted restriction fragment length polymorphism (RFLP) in

strain PH105 (aw), demonstrating the loss of a PstI site

located between the two BamHI restriction sites in the uncB

(a) gene (see Figure 3-1). The result demonstrated that

strain PH105 (a,) contains the 617 bp deletion from the

uncB (a) gene.

Comparison of Strain PH105 and Strain RH305

As demonstrated above strain PH105 (a,) and strain

RH305 (am,2) failed to grow on succinate minimal A media.

However, when the two strains were compared using

biochemical assays, an obvious difference between them

became apparent. There was a reduction in the specific

activity of ATP hydrolysis associated with membrane vesicles

isolated from the strain PH105 (ad). A similar level of

activity was also observed in strain PH105 (a,) carrying

plasmid pPH12 (a). Table 3-2 shows the specific activity

PH105 1100


4 1I I
345 6


1-,1 tQ-o F-b -I81


A.-a G-Y I O-R IC.-I

A.q I G.T i I C-

Figure 3-1. Southern analysis of strain PH105 (a,)
and its parent strain 1100. Panel A. Autoradiograph of
Southern blot. Panel B. Shows the arrangement of the unc
operon with its restriction map. The genomic DNA was
digested with Pstl. The probe was the Sall fragment
indicated by the bar. Key: P = Pstl; B = BamHI; S = Sall.


B P8
0 1

Table 3-2
Membrane Associated ATPase Activities

Strain/ ATPase specific activity"
Plasmid Mutation (Amol ATP/min/mg protein)

PH105 a, 0.40 .07
PH105/pPH12 a 0.48 .09

RH305 a,,, 1.05 .04
RH305/pPH12 a 0.92 .14

'ATPase specific activity is the average of at least
two separate experiments. The activity of strain RH305
(a, plasmidid pPH12 (a) is indicative of a wild-type
strain. ATP hydrolyzed was determined by measuring release
of inorganic phosphate over a time course. Assays were
performed at pH 9.1 (see Chapter 2).

for membrane bound ATP hydrolysis for strains PH105 (a,)

and strain RH305 (as,). Apparently, the reduction in

specific activity is due to a reduction in the amount of F,

ATPase present in strain PH105 (ad) as compared to strain

RH305 (aa,,).

This deficiency was further demonstrated in the

fluorescence assay for ATP driven proton translocation.

Figure 3-2 displays the ACMA fluorescence traces for NADH

driven and ATP membrane energization. The traces for strain

PH105 (a,) carrying plasmids pPH12 (a) and pPH4 (a), showed

fluorescence quenching approximately half the level of

pPH12(a) in strain RH305 (aas2) or strain PH105 (a.)

carrying the plasmid pEMS54 (a,c,b,6,a,7y,,e). Strain PH105

(a,) carrying the plasmid pEMS54 (a,c,b,6,a,y,f,e)

displayed twice as much ATP driven fluorescence quenching

activity as strains of PH105 (a.) carrying plasmids pPH12

(a) or pPH4 (a). This indicated that strain PH105 (a.)

carrying the plasmid pEMS54 (a,c,b,6,a,7,3,e) had

significantly higher levels of FIFo ATP synthase activity.

Strain RH305 (am,2) carrying plasmid pPH12 (a) had a

similar level of ACMA fluorescence quenching as strain PH105

(a,) carrying plasmid pEMS54 (a,c,b,8,a,7,#,e), indicating

comparable levels of FIFo ATP synthase. The difference

between strain RH305 (aEs,) carrying pPH12 (a) and strain

PH105 (aM) carrying pPH12 (a) indicated that strain PH105

(aM) was producing lower levels of other FIFo ATP synthase










2 min

Figure 3-2. Energization of membrane vesicles
determined by fluorescence quenching of ACMA. Reduced
fluorescence indicates establishment of a proton gradient
across the membrane via electron transport (NADH) or FFo
ATP synthase (ATP). ACMA was added to a final concentration
of 1 AM. Membrane vesicles were at a concentration of 250
Ag protein/ml of buffer (50 mM MOPS and 10 mM MgC12, pH
7.3). The reagents were added to final concentrations of:
NADH 0.5 mM; KCN 0.5 mM; ATP 0.4 mM; nigericin 0.05 iM.
the strain/plasmid combinations are listed under the traces
in the order of ATP driven fluorescence quenching levels.

subunits compared to strain RH305 (am,2). The NADH-driven

quenching data for strain PH105 (a,) carrying plasmid pPH12

(a) are also reduced indicating the membranes of this

particular preparation were leaky. This probably accounted

for the difference in ACMA fluorescence quenching observed

between strain PH105 (a,) carrying plasmid pPH12 (a) and

strain PH105 (a,) carrying plasmid pPH4 (a). The

manipulation of the promoter for the uncB (a) gene in

plasmid pPH12 (a) may have reduced the expression of the

uncB (a) gene (Chapter 2), but the high levels of ACMA

fluorescence quenching exhibited by strain RH305 (a2,y)

carrying pPH12 (a) indicate that this not the case.

Western Analysis of Strain PH105 and Strain RH305

To look directly at the expression of subunits of FIFo

ATP synthase, Western blots were done using anti-b

antibodies (Deckers-Hebestreit et al., 1992; McCormick et

al., 1993; Stack & Cain, 1994) and anti-FI antibodies

(Brusilow et al., 1981). Figure 3-3 shows the results of

the Western blot. A reduction in the amount of the b

subunit was seen in strain PH105 (a,) as compared to strain

RH305 (a.s). Whether plasmid pPH12 (a) was present or not

made no apparent difference to the amount of the b subunit

present in strain PH105 (a,). Strain RH305 (a.,,)

exhibited a reduced level of the b subunit present compared

to strain RH305 (a.,.) carrying plasmid pPH12 (a). The

lower molecular weight band observed on the Western may have

1100 RH305


pPH12 RH305 pPH12 PH105

Figure 3-3. Western analysis of strain RH305 (a),,y)
and strain PH105 (aM). Strain 1100, positive control.
Strain 1100ABC, negative control. Plasmid pPH12 (a). Panel
A: anti-b antibody. Panel B: anti-F1 antibody. See Chapter
2 for details.


1100 ABC





Figure 3-3--continued. Strain 1100, positive control.
Strain 1100ABC, negative control. Plasmid pPH12 (a). Panel
A: anti-b antibody. Panel B: anti-Fi antibody. See Chapter
2 for details.